” I think that it’s incredibly important for all programmers to have a public presence by being involved in open source or having side projects that are publicly available. The industry is quickly changing and more and more people are realizing how ineffective the standard techniques of programmer evaluation are. This includes things like resumes and coding questions “ –Nathan Marz.
Nathan Marz: You only live once, so it’s important to make the most of the time you have. I find Bezos’s “regret minimization framework” a great way to make decisions with a long term perspective. Too often people make decisions only thinking about marginal, short-term gains, and this can lead you down a path you never intended to go. And failure, if it happens, is not as bad as it seems. Worst comes to worse I’ll have learned an enormous amount, had a unique and interesting experience, and will just try something else.
Q2. Do you want to disclose in general terms what you’ll be working on?
Nathan Marz: Sorry, not at the moment. [ edit: as of now he did not disclose it]
Q3. You open-sourced Cascalog, ElephantDB, and Storm. Which of the three is in your opinion the most rewarding?
Nathan Marz: Storm has been very rewarding because of the sheer number of people using it and the diversity of industries it has penetrated, from healthcare to analytics to social networking to financial services and more.
Q4. What are in your opinion the current main challenges for Big Data analytics?
Nathan Marz: I think the biggest challenge is an educational one. There’s an overwhelming number of tools in the Big Data ecosystem, all very much different than the relational databases people are used to, and none is a one-sized-fits-all solution. This is why I’m writing my book “Big Data” – to show people a structured, principled approach to architecting data systems and how to use those principles to choose the right tool for you particular use case.
Q5. In January 2013, version 0.8.2 of Storm was released? What is new?
Nathan Marz: There was a lot of work done in 0.8.2 on making it easier to use a shared cluster for both production and in-development applications. This included improved monitoring support, which helps with detecting when you’ll need to scale with more resources, and a brand new scheduler that isolates production and development topologies from each other. And of course, the usual bug fixes and small improvements.
Q6. How do you expect Storm evolving?
Nathan Marz: There’s a lot of work happening right now on making Storm enterprise-ready. These include security features such as authentication and authorization, enhanced monitoring capabilities, and high availability for the Storm master. Long term, we want to continue with the theme of having Storm seamlessly integrate with your other realtime backend systems, such as databases, queues, and other services.
Q7. Daniel Abadi of Hadapt, said in a recent interview: “the prevalent architecture that people use to analyze structured and unstructured data is a two-system configuration, where Hadoop is used for processing the unstructured data and a relational database system is used for the structured data. However, this is a highly undesirable architecture, since now you have two systems to maintain, two systems where data may be stored, and if you want to do analysis involving data in both systems, you end up having to send data over the network which can be a major bottleneck.”
What is your opinion on this?
Nathan Marz: I think that “structured” vs. “unstructured” is a false dichotomy. It’s easy to store both unstructured and structured data in a distributed filesystem: just use a tool like Thrift to make your structured schema and serialize those records into files. A common objection to this is: “What if I need to delete or modify one of those records? You can’t cheaply do that when the data is stored in files on a DFS.” The answer is to move beyond the era of CRUD and embrace immutable data models where you only ever create or read data. In the architecture I’ve developed, which I call the “Lambda Architecture“, you then build views off of that data using tools like Hadoop and Storm, and it’s the views that are indexed and go on to feed the low latency requests in your system.
Q8. What are the main lessons learned in the last three years of your professional career?
Nathan Marz: I think that it’s incredibly important for all programmers to have a public presence by being involved in open source or having side projects that are publicly available. The industry is quickly changing and more and more people are realizing how ineffective the standard techniques of programmer evaluation are. This includes things like resumes and coding questions. These techniques frequently label strong people as weak or weak people as strong. Having good work out in the open makes it much easier to evaluate you as a strong programmer. This gives you many more job options and will likely drive up your salary as well because of the increased competition for your services.
For this reason, programmers should strongly prefer to work at companies that are very permissive about contributing to open source or releasing internal projects as open source. Ironically, a company having this policy is assisting in driving up the value (and price) of the employee, but as time goes on I think this policy will be necessary to even have access to the strongest programmers in the first place.
Nathan Marz. was the Lead Engineer at BackType before BackType was acquired by Twitter in July of 2011. At Twitter he started the streaming compute team which provides infrastructure that supports many critical applications throughout the company. He left Twitter in March of 2013 to start his own company (currently in stealth).
- Big Data: Principles and best practices of scalable realtime data systems.
Nathan Marz (Twitter) and James Warren
MEAP Began: January 2012
Softbound print: Fall 2013 | 425 pages
Download Chapter 1: A new paradigm for Big Data (.PDF)
follow ODBMS.org on Twitter: @odbmsorg
“Working with empirical genomic data and modern computational models, the laboratory addresses questions relevant to how genetics and the environment influence the frequency and severity of diseases in human populations” –Thibault de Malliard.
Big Data for Genomic Sequencing. On this subject, I have interviewed Thibault de Malliard, researcher at the University of Montreal’s Philip Awadalla Laboratory, who is working on bioinformatics solutions for next-generation genomic sequencing.
Q1. What are the main research activities of the University of Montreal’s Philip Awadalla Laboratory?
Thibault de Malliard: The Philip Awadalla Laboratory is the Medical and Population Genomics Laboratory at the University of Montreal. Working with empirical genomic data and modern computational models, the laboratory addresses questions relevant to how genetics and the environment influence the frequency and severity of diseases in human populations. Its research includes work relevant to all types of human diseases: genetic, immunological, infectious, chronic and cancer.
Using genomic data from single-nucleotide polymorphisms (SNP), next-generation re-sequencing, and gene expression, along with modern statistical tools, the lab is able to locate genome regions that are associated with disease pathology and virulence as well as study the mechanisms that cause the mutations.
Q2. What is the lab’s medical and population genomics research database?
Thibault de Malliard: The lab’s database is regrouping all the mutations (SNPs) found by DNA genotyping, DNA sequencing and RNA sequencing for each samples. There is also annotation data from public databases.
Q3. Why is data management important for the genomic research lab?
Thibault de Malliard: All the data we have is in text csv files. This is what our software takes as input, and it will output other text csv files. So we use a lot of Bash and Perl to extract the information we need and to do some stats. As time goes, we multiply the number of files by sample, by experiment, and finally we get statistics based on the whole data that need recalculating each time we perform a new sequencing/genotyping (mutation frequency, mutations per gene, etc).
With this database, we are also preparing for the lab’s future:
• As the amount of data increases, one day the memory will not fit an associative array.
• Looking to a 200 GB file to find one specific mutation will not be a good option.
• Adding new data to the current files will take more and more time/space.
• We need to be able to select the data according to every parameter we have, i.e., grouping by type of mutation and/or by chromosome, and/or by sample information by gender, ethnicity, age, or pathology.
• We then need to export a file, or count / sum / average it.
Q4. Could you give us a description of what kind of data is in the lab’s genomic research database storing and processing? And for what applications?
Thibault de Malliard: We are storing single nucleotide polymorphisms (SNPs), which are the most common form of genetic mutations among people, from sequencing and genotyping. When an SNP is found for a sample, we also look at what we have at the same position for the other samples:
• There is no SNP but data for the sample, so we know this sample does not have the SNP.
• There is no data for the sample, so we cannot assess whether or not there is an SNP for this sample at this position.
We gather between 1.8 and 2.5 million nucleotides (at least one sample has it) per sample, depending on the experiment technique. We store them in the database along with some information:
• how damaged the SNP can be for the function of the gene
• its frequency in different populations (African, European, French Canadian…).
The database also contains information about each sample, such as gender, ethnicity, pathology. This will keep growing with our needs. So, basically, we have a sample table, a mutations table with their information, an experiment table and a big table linking the 3 previous tables with relations one to many.
Here is a very slightly simplified example of a single record in our database:
|Type of data||Data||Table|
|SNP||T||Begin Mutation information table|
|Damaging for gene function?||synonymous|
|Present in known database?||yes||End Mutation information table|
|Sequencing quality||26||Begin Table linking other tables together||containing information about 1 mutation for 1 sample from 1 sequencing|
|Validated by another experiment?||no||End Table linking other tables|
|Sample||345||Begin Sample table|
|family||10||End Sample table|
|Sequencing information||Illumina Hiseq 2500||Begin Sequencing table|
|Sequencing type (DNA RNA…)||RNAseq|
|Analysis pipeline info||No PCR duplicates only Properly paired||End Sequencing table|
The applications are multiple, but here are some which come to my mind:
• extract subset of data to use with our tools
• doing stats, counts
• find specific data
• annotate our data with public databases
Q5. Why did you decide to deploy TokuDB database storage engine to optimize the lab’s medical and population genomics research database?
Thibault de Malliard: We knew that the data could not be managed with MySQL and MyISAM. One big issue is the insert rate, and TokuDB offered a solution up to 50 times faster. Furthermore, TokuDB allows us to manipulate the structure of the database without blocking access to it. As a research team, we always have new information to add, which means column additions.
Q6. Did you look/consider other vendor alternatives? If yes, which ones?
Thibault de Malliard: None. This is much too time consuming.
Q7. What are you specifically using TokuDB for?
Thibault de Malliard: We only store genetic data with information related to this genetic.
Q8. How many databases do you use? What are the data requirements?
Thibault de Malliard: I had planned to use three databases:
1. Database for RNA/DNA sequencing and from DNA genotyping (described before);
2. Database for data from well-known reference databases (dbsnp, 1000genome);
3. A last one to store analyzed data from database 1 and 2.
The data stored is manly the nucleotide (a character: A, C, G, T) with integer information like quality, position, and Boolean flags. I avoid using any string to keep the table as small as possible.
Q9. Especially, what are the requirements for data ingest of records and retrieve of data?
Thibault de Malliard: As a research team, we do not have high requirements like real-time insertion from logs. But I would say, at most, the import should take less than a night. The update of the database 1 is critical with the addition of a new sequencing or genotyping experiment: a batch of 50M records (can be more than 3 times higher!) has to be inserted. This has been happening monthly, but it should increase this year.
We have a huge amount of data, and we need to get query results as fast as possible, We have been used to one or two days (a weekend) of query time – having 10 seconds is much more preferable!
Q10. Could you give some examples of what are the typical research requests that need data ingestion and retrieval
Thibault de Malliard: We have a table with all the SNPs for 1000 samples. This is currently a 100GB table.
A typical query could be to get the sample that got a mutation different from the 999 others. We also have some samples that are families: a child with its parents. We want to find the SNPs present in this child, but not present in the other family member.
We may want to find mutations common to one group of sample given the gender, a disease state, ethnicity.
Q11. What kind of scalability problems did you have?
Thibault de Malliard: The problem is managing this huge amount of data. The number of connections should be very low. Most of the time, there is only one user. So I had to choose the data types carefully and the relationships between the tables. Lately, I ran into a very slow join with a range so I decided to split the position based tables by chromosome. Now there are 26 tables and some procedures to launch queries through the chromosomes. The gain of time is not quantifiable.
Q12. Do you have any benchmarking measures to sustain the claim that Tokutek’s TokuDB has improved scalability of your system?
Thibault de Malliard: I populated the database with two billion records in the main table and then did queries. While I did not see improvements with my particular workload for queries, I did see significant insertion performance gains. When I tried to add an extra 1M record (Load data infile), it took 51 minutes for MyISAM to load the data, but less than one minute with TokuDB. I extend this amount of data to an RNA sequencing experiment: it should take 2.5 days for MyISAM but one hour for TokuDB.
Q13. What are the lessons learned so far in using TokuDB database storage engine in your application domain?
Thibault de Malliard: We are still developing it and adding data. But inserting data into the two main tables (0.9G records, 2.3G records) was done in a fairly good time, less than one day. Adding columns to fulfill the needs of the team is also a very easy feature: it takes one second to create the column. Updating it is another story, but the table is still accessible during this process.
Another great feature, like the one I use with each query, is to be able to follow the state of the query.
You can follow in the process list the number of rows which were queried. So if you have a good estimation of the number of records expected, you know exactly the time of the query. I cannot count the number of process I killed because the query time expected was not acceptable.
Qx. Anything you wish to add?
Thibault de Malliard: The sequencing/genotyping technologies evolve very fast. Evolving means more data from the machines. I expect our data to grow at least three times each year. We are glad to have TokuDB in place to handle the challenge.
Since 2010, Thibault de Malliard has worked in the University of Montreal’s Philip Awadalla Laboratory where he provides bioinformatics support to the lab crew and develops bioinformatics solutions for next-generation genomic sequencing. Previously, he worked for the French National Institute for Agricultural Research (INRA) with the MIG laboratory (Mathematics, Informatics and Genomics) where, as part of the European Nanomubiop project, he was tasked with developing software to produce probes for a HPV chip. He holds a masters degree in bioinformatics (France).
- Big Data for Good. by Roger Barca, Laura Haas, Alon Halevy, Paul Miller, Roberto V. Zicari. June 5, 2012:
A distinguished panel of experts discuss how Big Data can be used to create Social Capital.
Blog Panel | Intermediate | English | DOWNLOAD (PDF)| June 2012|
Follow ODBMS.org on Twitter: @odbmsorg
“We were an early user of Hadoop and found ourselves pushing its scalability bounds and of necessity innovating our own solutions. For example we wrote our own API on top of it, rebuilt its sorter, and developed an alternative file system to achieve better performance and cost-effectiveness” — Jim Kelly and Sriram Rao.
Quantcast released last October the Quantcast File System (QFS) to open source. I have interviewed Jim Kelly, head of R&D at Quantcast, and Sriram Rao, Principal Scientist in Microsoft’s Cloud and Information Services Lab (CISL). Both worked closely on the development of the QFS platform at Quantcast.
Q1. What is the main business of Quantcast?
Jim Kelly: Quantcast measures digital audiences for millions of web destinations and helps advertisers reach their audiences more effectively. We observe billions of anonymous digital media consumption events every month and use large scale machine learning techniques to characterize audiences and to identify and reach relevant ones for advertisers.
We’re not in the distributed file system business, we’re in the same category as Google and Yahoo: an advertising company doing pioneering work in big data and sharing it with the rest of the community.
Q2. What are the main big data processing challenges you have experienced so far?
Jim Kelly: We’ve been doing big data processing since the launch of our measurement product in 2006, which put us in the big data space before it had become a space. We were an early user of Hadoop and found ourselves pushing its scalability bounds and of necessity innovating our own solutions. For example we wrote our own API on top of it, rebuilt its sorter, and developed an alternative file system to achieve better performance and cost-effectiveness. Our business and data volume have grown steadily since then. We currently receive over 50 TB of data per day, continually challenging us to operate at a scale beyond most organizations’ experience and most technologies’ comfort zone.
Sriram: In terms of big data, Hadoop is targeted to the sweet spot where the volume of data being processed in a computation is roughly the size of the amount of RAM in the cluster. At Quantcast, what we found was that we were frequently jobs where the amount of data being data processed per day (even with compression) was substantial. For instance, when I started at Quantcast in 2008, we were doing 100TB per day. Scaling the data processing volume required us to rethink components of Hadoop and build novel soutions. As we developed novel solutions and deployed them in our cluster, we were able to increase data processing volumes to as much as 500TB per day over a 2-year period. This increase occurred purely due to software improvements on the same hardware.
Q3. What lessons did you learn so far in using Hadoop for Big Data Analytics?
Jim Kelly: Hadoop has been a fantastic boon for us and the rest of the community. By providing a framework for developers (and increasingly, non-developers) to do parallel computing without a lot of systems knowledge, it has democratized an arcane and difficult problem. It has been a catalyst for the emergence of the big data community: a whole ecosystem of people, companies and technologies focused on different aspects of the problem. We’ve been a big beneficiary, both in terms of technology we’ve been able to use directly, and team members we’ve been able to attract because they want to do cutting-edge work in the dynamic space Hadoop has created. The lesson, I think, is how much value a technology can provide above and beyond what it does when you run it.
Sriram: The Apache Hadoop distribution is the same one that powers Yahoo!’s large clusters. Having an open source software tested at such scale has provided a great starting point for companies looking to process big data. While the data processing needs at Quantcast are extreme (10’s of PB per day), the main lesson I think is that, for a vast majority of the companies interested in data analytics Hadoop has become a stable platform which works “out-of-the-box”.
Q4. What are the scalability limits of current implementation of Hadoop?
Jim Kelly: It’s difficult to generalize, because there are so many different environments using Hadoop in different ways. What becomes an issue for one set of use cases on one set of hardware is not necessarily relevant for another, and I think it’s safe to say that the great majority of environments working with Hadoop needn’t worry about its scalability.
In our environment we’re processing a steady diet of many-terabyte jobs on many hundreds of servers, and we’ve invested a great deal in making the “sort” phase of MapReduce more efficient. In between the “map” and the “reduce,” Hadoop has to sort and group the mapper output and distribute it to the appropriate reducers. In general every reducer will have to read some of every mapper’s output, and with M mappers and N reducers that causes MxN disk seeks. When your jobs are large enough to require M and N in the thousands, this becomes a big performance bottleneck. We wrote a blog post describing the problem in more detail.
In our own environment we’ve achieved more efficiency at this scale by developing the Quantcast File System (QFS) and Quantsort. Sriram joined us in 2008 and led the effort to adapt the Kosmos File System (KFS) to the job and to build a new sort architecture on top of it. Using the two systems, we’ve sorted data sets of nearly a petabyte, and we regularly process over 20 PB per day.
QFS also improved where Hadoop was scaling too readily for us: cost. By using a more efficient data encoding, it lets us do more data processing using only half the hardware we would need with Hadoop’s HDFS, and therefore half the power and half the cooling and half the maintenance.
Sriram: The fundamental issue with Hadoop is how “shuffle” phase is handled as the data volume scales.
In a Map-Reduce computation, the output generated by map tasks during the “map” phase is partitioned and a reducer obtains its input partition from each of the map tasks. In particular, for large jobs, the size of the map output may exceed the amount of RAM in the cluster. When that happens, data has to be written to disk and then read back and transported over the network.
The “shuffle” is an “external” distributed merge-sort and is known to be seek intensive. Hadoop uses traditional mechanisms for implementing the shuffle and this is known to have scalability issues when the volume of data to be shuffled exceeds the amount of RAM in the cluster.
Q5. Sriram what was your contribution to the QFS project?
Sriram: QFS is an evolution of the KFS system that I started at Kosmix. The bulk of the KFS code base is in QFS. At Quantcast, we continued to evolve KFS by improving performance, by adding features, etc. One key feature we added at Quantcast was support multi-writer append (i.e., atomic append). It turns out that this feature can be leveraged to build high-performance sort architectures, which are fundamental to doing Map-Reduce computations at scale. In general, I worked on all aspects of the KFS system.
Q6. How does the Kosmos File System (KFS) relate to the QFS?
Jim Kelly: QFS evolved from KFS and has a couple years’ improvements on it. The most significant is Reed-Solomon erasure coding, which doubles effective storage and improves performance.
Sriram: The QFS project has its roots in the KFS system that I built. The bulk of the KFS code base is in QFS.
I started KFS originally at Kosmix Corp (now, @WalmartLabs) in 2006. At that point, Kosmix was trying to build a search engine and KFS was intended to be the storage substrate. The KFS architecture is based on the Google filesystem paper. The main idea in KFS is that blocks of a file are striped across nodes and replicated for fault-tolerance. We eventually released KFS as open source in 2007, after which I joined Quantcast. At Quantcast, we continued to evolve KFS by improving performance, by adding features, etc. KFS 0.5 is a snapshot of the code base that was run in production cluster at Quantcast in 2010. So, if you will, think of KFS as QFS 0.5. The significant difference between the two systems is erasure coding.
Q7. Sriram you also worked on another open source project Sailfish. Could you tell us a bit more about it?
Sriram: Data analytics computations are run on machines in a datacenter. The frameworks for doing these computations were designed in 2004-2006 timeframe when bandwidth within the datacenter was a scarce resource. However, that design point is changing.
Over the next few years, technological trends suggest that bandwidth within a datacenter will substantially increase. Inter-node connectivity is expected to go up from 1Gbps between any pair of nodes to 10Gbps (and possibly higher). Given such plentiful bandwidth, how would we build analytical engines (such as, Map-Reduce computation engines)? Sailfish is an attempt at re-designing how the shuffle data is transported in Map-Reduce frameworks by taking advantage of better connectivity.
Sailfish design is based on two key ideas:
1. Use network-wide data aggregation to improve disk subsystem performance.
2. Gather statistics on the aggregated data to plan for reduce phase of execution.
We instantiated these two ideas in the context of Hadoop-0.20.2, where we re-worked the entire shuffle pipeline and how the reduce phase of execution is done. Sailfish leverages the concurrent append capabilities of KFS to move data from mappers to reducers. When a Hadoop Map-Reduce job is run with Sailfish, (1) user does not tune sort parameters and (2) number of reducers in a job and their task assignment is determined at run-time in a data-dependent manner. Essentially, Sailfish simplifies running large-scale Hadoop-based Map-Reduce computations. We have also added support for preemption in Sailfish, where the preemption is implemented via a checkpoint/restart mechanism. This allows us to seamlessly handle skew.
Here is the Sailfish website which has links to the papers we have published and the software we have released.
Q8. Who is currently using Sailfish?
Sriram: Sailfish is a research prototype which has been released to open-source so that other researchers/users can benefit from our work. We are currently working on migrating many of the ideas from Sailfish to Hadoop version 2.0 (aka YARN). We have filed some JIRAs and intend to contribute the code to Hadoop.
Q9. Quantcast released last October the Quantcast File System (QFS) to open source, (link here). Why?
Jim Kelly: We’ve always relied heavily on open source software internally, and it’s good to be able to give back. It’s a chance for developers to have more of an impact than they could working on a proprietary system. And we believe file systems particularly, as foundational infrastructure components, benefit from the open source model where they can get a lot of scrutiny and be challenged to prove themselves in many different environments.
Q10. Is QFS supposed to be alternative to Apache Hadoop`s HDFS or just a complementary solution?
Jim Kelly: Hadoop has a wide variety of customers with diverse priorities. Some are just getting started and put a lot of value on ease of use. Others care most about performance at small scale, or high availability. HDFS offers broad functionality to meet needs for all of them.
Quantcast’s priorities are cost efficiency and high performance at large scale, so in developing QFS we went deeper around that use case. QFS has significantly improved the performance and economics of our data processing and can do the same for other organizations running their own clusters at large scale. By contrast organizations just getting started or needing specific HDFS features will probably find HDFS is a better fit.
Q11. What are the key technical innovations of QFS?
Jim Kelly: Architecturally QFS places a couple of different bets than HDFS. It’s implemented in C++ rather than Java, which allows better control and optimization in a number of areas. It also leans more heavily on modern networks, which have become much faster since HDFS launched and allow different optimization choices.
Those differences in approach make possible QFS’s Reed-Solomon erasure coding, which doubles storage efficiency and improves performance. The concurrent append feature allows many processes to write simultaneously to a single file, enabling applications like Quantsort. QFS’s use of direct I/O also improves both performance and manageability, by ensuring processes live within a predictable memory footprint.
Sriram: One of KFS’s key innovations that we did at Quantcast was support for “atomic append”.
This is the ability to allow multiple writers to append data to a file. Basically, think of the file as a “bag” into which data is dropped and there are no ordering guarantees on the order in which data is appended to a file. It turned out that this feature can be leveraged to build high-performance sort engines.
Q12. What kind of benchmark results do you have comparing QFS with Hadoop`s HDFS? What benchmark did you use for that?
Jim Kelly: Both QFS and HDFS rely on a central server to manage the file system, and its performance is critical as every cluster operation is dispatched through it. We’ve tested the two head-to-head for various operations and found QFS’s leaner C++ implementation pays off. Directory listings were 22% faster, and directory creation was nearly three times as fast.
These were our tests, and we encourage other people to run their own. We’ve put scripts and instructions to help on github.
Our whole-cluster tests will be harder for others to replicate at the same scale we did on our production cluster. We attempted to get some sense of how the file system affects real-world job run time. We measured end-to-end run time for simple Hadoop jobs that wrote or read 20TB of user data on an otherwise idle cluster, using QFS or HDFS as the underlying file system. The average throughput of the write job was about 75% higher using QFS, due to having to write less physical data. The read job throughput was 47% higher, primarily due to better parallelism. More information about these tests is available on the QFS wiki.
Q13. How reliable is QFS? What is your experience in using QFS in the Quantcast environment?
Jim Kelly: We started running KFS internally in 2008 on secondary storage and over years evolved it into today’s QFS, in the process adding features to make it more manageable at large scale and hardening it for production use. By 2011 it was running reliably enough that we were comfortable going all in. We copied all our data over from HDFS and have been running QFS exclusively since then for all our storage and MapReduce work. QFS has handled over six exabytes of I/O since then, so we’re confident it’s ready for other organizations’ production workloads.
Q14. You mentioned Hadoop’s architectural limitations for large jobs (terabytes and beyond). The more tasks a job uses, the less efficient its disk I/O becomes. To improve that, Quantcast developed an improved Hadoop Sort, called Quantsort. What is special about Quantsort?
Jim Kelly: Quantsort avoids the MxN combinatorics Hadoop faces shuffling data between mappers and reducers. It leverages QFS’s concurrent append feature to shuffle data through a fixed number of intermediate files. Reducers read their data in larger chunks, keeping disk I/O efficient. We wrote a blog post describing Quantsort more fully.
Q15. How is the performance of Quantsort?
Jim Kelly: Fast. Sriram did Quantcast’s first large sort back in 2009, of 140 TB in 6.2 hours on much less hardware than we have today. Quantsort’s record on a production job is 971 TB in 6.4 hours.
Sriram: For large jobs (think 10’s-100’s of TB of data and beyond) where the map output exceeds the size of RAM, Quantsort performance can be upto an order of magnitude faster when compared to Hadoop’s shuffle. Quantsort enables you to process more data quickly using the same hardware. Since jobs finish faster, it has the multiplier effect of allowing users to run more jobs on the cluster.
Q17. Is Quantsort part of QFS and therefore, also open source?
Jim Kelly: No, Quantsort is a separate code base.
Q18 In 2011, Quantcast migrated all production data to QSF and turned off HDFS. How did you manage to migrate the data stored in HDFS to QSF?
Jim Kelly: The data migration was very straightforward, a matter of running a trivial MapReduce job that copies data from one place to another. We avoided any outage for our production jobs by phasing them over, making them read from either file system but write to QFS during the transition.
Q19. How easy is to plug-in and use QFS into an existing Hadoop cluster which has already data stored in HDFS?
Jim Kelly: QFS is plug-in compatible with Hadoop, using KFS bindings that are already included in standard Hadoop distributions, so it’s very easy to integrate. The data storage format is different, requiring either migrating data to QFS or phasing it in gradually. The QFS wiki has a migration guide.
Q20. KFS has also been released an open source project. How does it relate to the Quantcast File System (QFS)?
Jim Kelly: Quantcast adopted KFS in 2008, became the main sponsor of contributions to the KFS project for the next two years and subsequently evolved it into QFS. You’ll find the concurrent append feature in the KFS repository, for example, but Reed-Solomon encoding is only in QFS.
Qx Anything you wish to add?
Jim Kelly: An invitation to other organizations doing distributed batch processing and interested in doubling their cluster capacity to give QFS a try. We’re happy to field questions or receive feedback on the QFS mailing list.
Jim Kelly leads Quantcast’s R&D team, which works on both adding computing capacity through cluster software innovations and using it up through new analytic and modeling products. Having been at Quantcast for six years, he has seen its data volumes and processing challenges grow from zero to petabytes and led technical and organizational changes that have kept Quantcast a step ahead. Previously he held engineering leadership roles at Oracle, Kana, and Scopus Technology (acquired by Siebel). Jim holds a PhD in physics from Princeton University.
Sriram Rao is currently a Principal Scientist in Microsoft’s Cloud and Information Services Lab (CISL), and formerly lead the Quantcast’s Cluster Team. Sriram was the creator of the Kosmos File System (KFS), which would become the underlying architecture for the Quantcast File System (QFS) that was released to open source last year. Sriram kiis also the creator of Sailfish, another open source project that is geared towards processing massive amounts of data. Previously he has worked at Quantcast, Yahoo, and Kosmix (now @WalmartLabs). Sriram holds a PhD in Computer Sciences from UT-Austin.
-ODBMS.org: Big Data and Analytical Data Platforms – Free Software
-ODBMS.org: Big Data and Analytical Data Platforms – Articles
Follow ODBMS.org on Twitter: @odbmsorg
“So the synergies in data management come not from how the systems connect but how the data is used to derive business value” –Steve Shine,
On Dec. 21, 2012, Actian Corp. announced the completion of the transaction to buy Versant Corporation. I have interviewed Steve Shine, CEO and President, Actian Corporation.
Q1. Why acquiring an object-oriented database company such as Versant?
Steve Shine: Versant Corporation, like us, has a long pedigree in solving complex data management in some of the world’s largest organisations. We see many synergies in bringing the two companies together. The most important of these is together we are able to invest more resources in helping our customers extract even more value from their data. Our direct clients will have a larger product portfolio to choose from, our partners will be able to expand in adjacent solution segments, and strategically we arm ourselves with the skills and technology to fulfil our plans to deliver innovative solutions in the emerging Big Data Market.
Q2. For the enterprise market, Actian offers its legacy Ingres relational database. Versant on the other hand offers an object oriented database, especially suited for complex science/engineering applications. How does this fit? Do you have a strategy on how to offer a set of support processes and related tools for the enterprise? if yes, how?
Steve Shine: While the two databases may not have a direct logical connection at client installations, we recognise that most clients use these two products as part of a larger more holistic solutions to support their operations. The data they manage is the same and interacts to solve business issues – for example object stores to manage the relationships between entities; transactional systems to manage clients and the supply chain and analytic systems to monitor and tune operational performance. – Different systems using the same underlying data to drive a complex business.
We plan to announce a vision of an integrated platform designed to help our clients manage all their data and their complex interactions, both internal and external so they can not only focus on their running their business, but better exploit the incremental opportunity promised by Big Data.
Q3. Bernhard Woebker, president and chief executive officer of Versant stated, “the combination of Actian and Versant provides numerous synergies for data management”. Could you give us some specific examples of such synergies for data management?
Steve Shine: Here is a specific example of what I mean by helping clients extracting more value from data in the Telco space. These type of incremental opportunities exist in every vertical we have looked at.
An OSS system in a Telco today may use an Object store to manage the complex relationships between the data, the same data is used in a relational store to monitor, control and manage the telephone network.
Another relational store using variants of the same data manages the provisioning, billing and support for the users of the network. The whole data set in Analytical stores is used to monitor and optimise performance and usage of the network.
Fast forwarding to today, the same data used in more sophisticated ways has allowed voice and data networks to converge to provide a seamless interface to mobile users. As a result, Telcos have tremendous incremental revenue opportunities BUT only if they can exploit the data they already have in their networks. For example: The data on their networks has allowed for a huge increase in location based services, knowledge and analysis of the data content has allowed providers to push targeted advertising and other revenue earning services at their users; then turning the phone into a common billing device to get even a greater share of the service providers revenue… You get the picture.
Now imagine other corporations being able to exploit their information in similar ways: Would a retailer benefit from knowing the preferences of who’s in their stores? Would a hedge fund benefit from detecting a sentiment shift for a stock as it happens? Even knowledge of simple events can help organisations become more efficient.
A salesman knowing immediately a key client raises a support ticket; A product manager knowing what’s being asked on discussion forums; A marketing manager knowing a perfect prospect is on the website.
So the synergies in data management come not from how the systems connect but how the data is used to derive business value. We want to help manage all the data in our customers organisations and help them drive incremental value from it. That is the what we mean by numerous synergies from data management and we have a vision to deliver it to our customers.
Q4. Actian claims to have more than 10,000 customers worldwide. What is the value proposition of Versant’s acquisition for the existing Actian`s customers?
Steve Shine: I have covered this in the answers above. They get access to a larger portfolio of products and services and we together drive a vision to help them extract greater value from their data.
Q5. Versant claims to have more than 150,000 installations worldwide. How do you intend to support them?
Steve Shine: Actian already runs a 24/7 global support organisation that prides itself in delivering one of the industry’s best client satisfaction scores. As far as numbers are concerned, Versant’s large user count is in essence driven by only 250 or so very sophisticated large installations whereas Actian already deals with over 10,000 discreet mission critical installations worldwide. So we are confident of maintaining our very high support levels and the Versant support infrastructure is being integrated into Actian’s as we speak.
Q6. Actian is active in the market for big data analytics. How does Versant’s database technology fit into Actian’s big data analytics offerings and capabilities?
Steve Shine: Using the example above imagine using OSS data to analyse network utilisation, CDR’s and billing information to identify pay plans for your most profitable clients.
Now give these clients the ability to take business action on real time changes in their data.Now imagine being able to do that from an integrated product set from one vendor. We will be announcing the vision behind this strategy this quarter. In addition, the Versant technology gives us additional options for solutions for big data for example visualisation and managing meta data.
Q7. Do you intend to combine or integrate your analytics database Vectorwise with Versant’s database technology (such as Versant JPA)? If yes, how?
Steve Shine: Specific plans for integrating products within the overall architecture have not been formulated. We have a strong philosophy that you should use the best tool for the job eg OODB for some things, OLTP RDBMS for other etc. But the real value comes from being able to perform sophisticated analysis and management across the different data stores. That is part of the work out platform integration efforts are focused on.
Q8. What are the plans for future software developments. Will you have a joint development team or else?
Steve Shine: We will be merging the engineering teams to focus on providing innovative solutions for big Data under single leadership.
Q9. You have recently announced two partnerships for Vectorwise, with Inferenda and BiBoard. Will you also pursue this indirect channel path also for Versant’s database technology?
Steve Shine: The beauty of the vision we speak of is that our joint partner have a real opportunity to expand their solutions using Actian’s broader product set and for those that are innovative the opportunity for new emerging markets
Q10. Versant recently developed Versant JPA. Is the Java market important for Actian?
Steve Shine: Yes !
Q11. It is currently a crowded database market: several new database vendors (NoSQL and NewSQL) offering innovative database technology (NuoDB, VoltDB, MongoDB, Cassandra, Couchbase, Riak to name a few), and large companies such as IBM and Oracle, are all chasing the big data market. What is your plan to stand out of the crowd?
Steve Shine: We are very excited about the upcoming announcement on our plans for the Big Data market. We will be happy to brief you on the details closer to the time but I will say that early feedback from analysts houses like Gartner have confirmed that our solution is very effective and differentiated in helping corporations extract business value from Big Data. On a higher scale, many of the start ups are going to get a very rude awakening when they find that delivering a database for mission critical use is much more than speed and scale of technology. Enterprises want world class 24×7 support service with failsafe resilience and security. Real industry grade databases take years and many $m’s to reach scalable maturity. Most of the start ups will not make it. Actian is uniquely positioned in being profitable and having delivered industry grade database innovation but also being singularly focused around data management unlike the broad, cumbersome and expensive bigger players. We believe value conscious enterprises will see our maturity and agility as a great strength.
Qx Anything else you wish to add?
Steve Shine: DATA! – What a great thing to be involved in! Endless value, endless opportunities for innovation and no end in sight as far as growth is concerned. I look forward to the next 5 years.
Steve Shine, CEO and President, Actian Corporation.
Steve comes to Actian from Sybase where he was senior vice president and general manager for EMEA, overseeing all operational, sales, financial and human resources in the region for the past three years. While at Sybase, he achieved more than 200 million in revenue and managed 500 employees, charting over 50 percent growth in the Business Intelligence market for Sybase. Prior to Sybase, Steve was at Canadian-based Geac Computer Corporation for ten successful years, helping to successfully turn around two major global divisions for the ERP firm.
Follow ODBMS.org on Twitter: @odbmsorg
“The data you’re likely to need for any real-world predictive model today is unlikely to be sitting in any one data management system. A data scientist will often combine transactional data from a NoSQL system, demographic data from a RDBMS, unstructured data from Hadoop, and social data from a streaming API” –David Smith.
On the subject of Big Data Analytics I have interviewed David Smith, Vice President of Marketing and Community at Revolution Analytics.
Q1. How would you define the job of a data scientist?
David Smith: A data scientist is someone charged of analyzing and communicating insight from data.
It’s someone with a combination of skills: computer science, to be able to access and manipulate the data; statistical modeling, to be able to make predictions from the data; and domain expertise, to be able to understand and answer the question being asked.
Q2. What are the main technical challenges for Big Data predictive analytics?
David Smith: For a skilled data scientist, the main challenge is time. Big Data takes a long time just to move (so don’t do that, if you don’t have to!), not to mention the time required to apply complex statistical algorithms. That’s why it’s important to have software that can make use of modern data architectures to fit predictive models to Big Data in the shortest time possible. The more iterations a data scientist can make to improve the model, the more robust and accurate it will be.
Q3. R is an open source programming language for statistical analysis. Is R useful for Big Data as well? Can you analyze petabytes of data with R, and at the same time ensure scalability and performance?
David Smith: Petabytes? That’s a heck of a lot of data: even Facebook has “only” 70 Pb of data, total. The important thing to remember is that “Big Data” means different things in different contexts: while raw data in Hadoop may be measured in the petabytes, by the time a data scientist selects, filters and processes it you’re more likely to be in the terabytes or even gigabyte range when the data’s ready to be applied to predictive models.
Open Source R , with its in-memory, single-threaded engine, will still struggle even at this scale, though. That’s why Revolution Analytics added scalable, parallelized algorithms to R, making predictive modeling on terabytes of data possible. With Revolution R Enterprise , you can use SMP servers or MPP grids to fit powerful predictive models to hundreds of millions of rows of data in just minutes.
Q4. Could you give us some information on how Google, and Bank of America use R for their statistical analysis?
David Smith: Google has more than 500 R users , where R is used to study the effectiveness of ads, for forecasting, and for statistical modeling with Big Data.
In the financial sector, R is used by banks like Bank of America and Northern Trust and insurance companies like Allstate for a variety of applications, including data visualization, simulation, portfolio optimization, and time series forecasting.
Q5. How do you handle the Big Data Analytics “process” challenges with deriving insight?
- capturing data
- aligning data from different sources (e.g., resolving when two objects are the same)
- transforming the data into a form suitable for analysis
- modeling it, whether mathematically, or through some form of simulation
- understanding the output
- visualizing and sharing the results
David Smith: These steps reflect the fact that data science is an iterative process: long gone are the days where we would simply pump data through a black-box algorithm and hope for the best. Data transformation, evaluation of multiple model options, and visualizing the results are essential to creating a powerful and reliable statistical model. That’s why the R language has proven so popular: its interactive language encourages exploration, refinement and presentation, and Revolution R Enterprise provides the speed and big-data support to allow the data scientist to iterate through this process quickly.
Q6. What is the tradeoff between Accuracy and Speed that you usually need to make with Big Data?
David Smith: Real-time predictive analytics with Big Data are certainly possible. (See here for a detailed explanation.) Accuracy comes with real-time scoring of the model, which is dependent on a data scientist building the predictive model on Big Data. To maintain accuracy, that model will need to be refreshed on a regular basis using the latest data available.
Q7. In your opinion, is there a technology which is best suited to build Analytics Platform? RDBMS, or instead non relational database technology, such as for example columnar database engine? Else?
David Smith: The data you’re likely to need for any real-world predictive model today is unlikely to be sitting in any one data management system. A data scientist will often combine transactional data from a NoSQL system, demographic data from a RDBMS, unstructured data from Hadoop, and social data from a streaming API.
That’s one of the reasons the R language is so powerful: it provides interfaces to a variety of data storage and processing systems, instead of being dependent on any one technology.
Q8. Cloud computing: What role does it play with Analytics? What are the main differences between Ground vs Cloud analytics?
David Smith: Cloud computing can be a cost-effective platform for the Big-Data computations inherent in predictive modeling: if you occasionally need a 40-node grid to fit a big predictive model, it’s convenient to be able to spin one up at will. The big catch is in the data: if your data is already in the cloud you’re golden, but if it lives in a ground-based data center it’s going to be expensive (in time *and* money) to move it to the cloud.
David Smith, Vice President, Marketing & Community, Revolution Analytics
David Smith has a long history with the R and statistics communities. After graduating with a degree in Statistics from the University of Adelaide, South Australia, he spent four years researching statistical methodology at Lancaster University in the United Kingdom, where he also developed a number of packages for the S-PLUS statistical modeling environment.
He continued his association with S-PLUS at Insightful (now TIBCO Spotfire) overseeing the product management of S-PLUS and other statistical and data mining products. David smith is the co-author (with Bill Venables) of the popular tutorial manual, An Introduction to R, and one of the originating developers of the ESS: Emacs Speaks Statistics project.
Today, David leads marketing for REvolution R, supports R communities worldwide, and is responsible for the Revolutions blog.
Prior to joining Revolution Analytics, David served as vice president of product management at Zynchros, Inc.
Follow ODBMS.org on Twitter: @odbmsorg
“Our experience with MongoDB is that it’s architecture requires nodes to be declared as Master, and others as Slaves, and the configuration is complex and unintuitive. C* architecture is much simpler. Every node is a peer in a ring and replication is handled internally by Cassandra based on your desired redundancy level. There is much less manual intervention, which allows us to then easily automate many tasks when managing a C* cluster.” – Christos Kalantzis and Jason Brown.
Netflix, Inc. (NASDAQ: NFLX) is an online DVD and Blu-Ray movie retailer offering streaming movies through video game consoles, Apple TV, TiVo and more.
Last year, Netflix’s had a total of 29.4 million subscribers worldwide for their streaming service (Source).
I have interviewed Christos Kalantzis , Engineering Manager – Cloud Persistence Engineering and Jason Brown, Senior Software Engineer both at Netflix. They were involved in deploying Cassandra in a production EC2 environment at Netflix.
Q1. What are the main technical challenges that Big data analytics pose to modern Data Centers?
Kalantzis, Brown: As companies are learning how to extract value from the data they already have, they are also identifying all the value they could be getting from data they are currently not collecting.
This is creating an appetite for more data collection. This appetite for more data, is pushing the boundaries of traditional RDBMS systems and forcing companies to research alternative data stores.
This new data size, also requires companies to think about the extra costs involved storing this new data (space/power/hardware/redundant copies).
Q2. How do you handle large volume of data? (terabytes to petabytes of data)?
Kalantzis, Brown: Netflix does not have its own datacenter. We store all of our data and applications on Amazon’s AWS. This allows us to focus on creating really good applications without the overhead of thinking about how we are going to architect the Datacenter to hold all of this information.
Economy of Scale also allows us to negotiate a good price with Amazon.
Q3. Why did you choose Apache Cassandra (C*)?
Kalantzis, Brown: There’s several reasons we selected Cassandra. First, as Netflix is growing internationally, a solid multi-datacenter story is important to us. Configurable replication and consistency, as well as resiliency in the face of failure is an absolute requirement, and we have tested those capabilities more than once in production! Other compelling qualities include being an open source, Apache project and having an active and vibrant user community.
Q4: What do you exactly mean with “a solid multi-datacenter story is important to us”? Please explain.
Kalantzis, Brown: As we expand internationally we are moving and standing up new datacenters close to our new customers. In many cases we need a copy of the full dataset of an application. It is important that our Database Product be able to replicate across multiple datacenters, reliably, efficiently and with very little lag.
Q5. Tell us a little bit about the application powered by C*
Kalantzis, Brown: Almost everything we run in the cloud (which is almost the entire Netflix infrastructure) uses C* as a database. From customer/subscriber information, to movie metadata, to monitoring stats, it’s all hosted in Cassandra.
In most of our uses, Cassandra is the source of truth database. There are a few legacy datasets in other solutions, but are actively being migrated.
Q6: What are the typical data insights you obtained by analyzing all of these data? Please give some examples. How do you technically analyze the data? And by the way, how large are your data sets?
Kalantzis, Brown: All the data Netflix gathers goes towards improving the customer experience. We analyze our data to understand viewing preferences, give great recommendations and make appropriate choices when buying new content.
Our BI team has done a great job with the Hadoop platform and has been able to extract the information we need from the terabytes of data we capture and store.
Q7. What Availability expectations do you have from customers?
Kalantzis, Brown: Our internal teams are incredibly demanding on every part of the infrastructure, and databases are certainly no exception. Thus a database solution must have low latency, high throughput, strict uptime/availability requirements, and be scalable to massive amounts of data. The solution must, of course, be able to withstand failure and not fall over.
Q8: Be scalable up to?
Kalantzis, Brown: We don’t have an upper boundary to how much data an application can store. That being said, we also expect the application designers to be intelligent with how much data they need readily available in their OLTP system. Our Applications store anywhere from 10 GB to 100 terabyte [Edit corrected a typo] of data in their respective C* clusters. C* architecture is such that the cluster’s capacity grows linearly with every node added to the cluster. So in theory we can scale “infinitely”.
Q9. What other methods did you consider for continuous availability?
Kalantzis, Brown: We considered and experimented with MongoDB, yet the operational overhead and complexity made it unmanageable so we quickly backed away from it. One team even built a sharded RDBMS cluster, with every node in the cluster being replicated twice. This solution is also very complex to manage. We are currently working to migrate to C* for that application.
Q10. Could you please explain in a bit more detail, what kind of complexity made it unmanageable using MongoDB for you?
Kalantzis, Brown: Netflix strives to choose architectures that are simple and require very little in the way of manual intervention to manage and scale. Our experience with MongoDB is that it’s architecture requires nodes to be declared as Master, and others as Slaves, and the configuration is complex and unintuitive. C* architecture is much simpler. Every node is a peer in a ring and replication is handled internally by Cassandra based on your desired redundancy level. There is much less manual intervention, which allows us to then easily automate many tasks when managing a C* cluster.
Q11. How many data centers are you replicating among?
Kalantzis, Brown: For most workloads, we use two Amazon EC2 regions. For very specific workloads, up to four EC2 regions.
To slice that further, each region has two to three availability zones (you can think of an availability zone as the closest equivalent to a traditional data center). We shard and replicate the data within a region across it’s availability zones, and replicate between regions.
Q12: When you shard data, don’t you have a possible data consistency problem when updating the data?
Kalantzis, Brown: Yes, when writing across multiple nodes there is always the issue of consistency. C* “solves” this by allowing the client (Application) to choose the consistency level it desires when writing. You can choose to write to only 1 node, all nodes or a quorum of nodes. Each choice offers a different level of consistency and the application will only return from its write statement when the desired consistency is reached.
The same is with reading, you can choose the level of consistency you desire with every statement. The timestamp of each record will be compared, and make sure to only return the latest record.
In the end the application developer needs to understand the trade offs of each setting (speed vs consistency) and make the appropriate decision that best fits their use case.
Q13. How do you reduce the I/O bandwidth requirements for big data analytics workloads?
Kalantzis, Brown: What is great about C*, is that it allows you to scale linearly with commodity hardware. Furthermore adding more hardware is very simple and the cluster will rebalance itself. To solve the I/O issue we simply add more nodes. This reduces the amount of data stored in each node allowing us to get as close as possible to the ideal data:memory ratio of 1.
Q14. What is the tradeoff between Accuracy and Speed that you usually need to make with Big Data?
Kalantzis, Brown: When we chose C* we made a conscious decision to accept eventually consistent data.
We decided write/read speed and high availability was more important than consistency. Our application is such that this tradeoff, although might rarely provide inaccurate data (start a movie at the wrong location), it does not negatively impact a user. When an application does require accuracy, then we increase the consistency to quorum.
Q15. How do you ensure that your system does not becomes unavailable?
Kalantzis, Brown: Minimally, we run our c* clusters (over 50 in production now) across multiple availabilty zones within each region of EC2. If a cluster is multiregion (multi-datacenter), then one region is naturally isolated (for reads) from the other; all writes will eventually make it to all regions due to cass
Q16. How do you handle information security?
Kalantzis, Brown: We currently rely on Amazon’s security features to limit who can read/write to our C* clusters. As we move forward with our cloud strategy and want to store Financial data in C* and in the cloud, we are hoping that new security features in DSE 3.0 will provide a roadmap for us. This will enable us to move even more sensitive information to the Cloud & C*.
Q17. How is your experience with virtualization technologies?
Kalantzis, Brown: Netflix runs all of its infrastructure in the cloud (AWS). We have extensive experience with Virtualization, and fully appreciate all the Pros and Cons of virtualization.
Q18 How operationally complex is it to manage a multi-datacenter environment?
Kalantzis, Brown: Netflix invested heavily in creating tools to manage multi-datacenter and multi-region computing environments. Tools such as Asgard & Priam, has made that management easy and scalable.
Q19. How do you handle modularity and flexibility?
Kalantzis, Brown: At the data level, each service typically has it’s own Cassandra cluster so it can scale independently of other services. We are developing internal tools and processes for automatically consolidating and splitting clusters to optimize efficiency and cost. At the code level, we have built and open sourced our java Cassandra client, Astyanax , and added many recipes for extending the use patterns of Cassandra.
Q20. How do you reduce (if any) energy use?
Kalantzis, Brown: Currently, we do not. However, C* is introducing a new feature in version 1.2 called ‘virtual nodes’. The short explanation of virtual nodes is that it makes scaling up and down the size of a c* cluster much easier to manage. Thus, we are planning on using the virtual nodes concept with EC2 auto-scaling, so we would scale up the number of nodes in a C* cluster during peak times, and reduce the node count during troughs. We’ll save energy (and money) by simply using less resources when demand is not present.
Q21. What advice would you give to someone needing to ensure continuous availability?
Kalantzis, Brown: Availability is just one of the 3 dimensions of CAP (Concurrency, Availability and “Partitionability”). They should evaluate which of the other 2 dimensions are important to them and choose their technology accordingly. C* solves for A&P (and some of C).
Q22. What are the main lessons learned at Netflix in using and deploying Cassandra in a production EC2 environment?
Kalantzis, Brown: We learned that deploying C* (or any database product) in a virtualized environment has trade offs. You trade the fact that you have easy and quick access to many virtual machines, yet each virtual machine has MUCH less IOPS than a traditional server. Netflix has learned how to deal with those limitations and trade-offs, by sizing clusters appropriately (number of nodes to use).
It has also forced us to reimagine the DBA role as a DB Engineer role, which means we work closely with application developers to make sure that their schema and application design is as efficient as possible and not access the data store unnecessarily.
(@chriskalan) Engineering Manager – Cloud Persistence Engineering Netflix
Previously Engineering-Platform Manager YouSendIt
A Tech enthusiast at heart, I try to focus my efforts in creating technology that enhances our lives.
I have built and lead teams at YouSendIt and Netflix which has lead to the scaling out of persistence layers, the creation of a cloud file system and the adoption of Apache Cassandra as a scalable and highly available data solution.
I’ve worked as a DB2, SQL Server and MySQL DBA for over 10 years and through, sometimes painful, trial and error I have learned the advantages and limitations of RDBMS and when the modern NoSQL solutions make sense.
I believe in sharing knowledge, that is why I am a huge advocate of Open Source software. I share my software experience through blogging, pod-casting and mentoring new start-ups. I sit on the tech advisory board of the OpenFund Project which is an Angel VC for European start-ups.
(@jasobrown), Senior Software Engineer, Netflix
Jason Brown is a Senior Software Engineer at Netflix where he led the pilot project for using and deploying Cassandra in a production EC2 environment. Lately he’s been contributing to various open source projects around the Cassandra ecosystem, including the Apache Cassandra project itself. Holding a Master’s Degree in Music Composition, Jason longs to write a second string quartet.
Follow ODBMS.org on Twitter: @odbmsorg
“With MySQL 5.6, developers can now commingle the “best of both worlds” with fast key-value look up operations and complex SQL queries to meet user and application specific requirements” –Tomas Ulin.
On February 5, 2013, Oracle announced the general availability of MySQL 5.6.
I have interviewed Tomas Ulin, Vice President for the MySQL Engineering team at Oracle. I asked him several questions on the state of the union for MySQL.
Q1. You support several different versions of the MySQL database. Why? How do they differ with each other?
Tomas Ulin: Oracle provides technical support for several versions of the MySQL database to allow our users to maximize their investments in MySQL. Additional details about Oracle’s Lifetime Support policy can be found here.
Each new version of MySQL has added new functionality and improved the user experience. Oracle just made available MySQL 5.6, delivering enhanced linear scalability, simplified query development, better transactional throughput and application availability, flexible NoSQL access, improved replication and enhanced instrumentation.
Q2. Could you please explain in some more details how MySQL can offer a NoSQL access?
Tomas Ulin: MySQL 5.6 provides simple, key-value interaction with InnoDB data via the familiar Memcached API. Implemented via a new Memcached daemon plug-in to mysqld, the new Memcached protocol is mapped directly to the native InnoDB API and enables developers to use existing Memcached clients to bypass the expense of query parsing and go directly to InnoDB data for lookups and transactional compliant updates. With MySQL 5.6, developers can now commingle the “best of both worlds” with fast key-value look up operations and complex SQL queries to meet user and application specific requirements. More information is available here.
MySQL Cluster presents multiple interfaces to the database, also providing the option to bypass the SQL layer entirely for native, blazing fast access to the tables. Each of the SQL and NoSQL APIs can be used simultaneously, across the same data set. NoSQL APIs for MySQL Cluster include memcached as well as the native C++ NDB API, Java (ClusterJ and ClusterJPA) and HTTP/REST. Additionally, during our MySQL Connect Conference last fall, we announced a new Node.js NoSQL API to MySQL Cluster as an early access feature. More information is available here.
Q3. No single data store is best for all uses. What are the applications that are best suited for MySQL, and which ones are not?
Tomas Ulin: MySQL is the leading open source database for web, mobile and social applications, delivered either on-premise or in the cloud. MySQL is increasingly used for Software-as-a-Service applications as well.
It is also a popular choice as an embedded database with over 3,000 ISVs and OEMs using it.
MySQL is also widely deployed for custom IT and departmental enterprise applications, where it is often complementary to Oracle Database deployments. It also represents a compelling alternative to Microsoft SQL Server with the ability to reduce database TCO by up to 90 percent.
From a developer’s perspective, there is a need to address growing data volumes, and very high data ingestion and query speeds, while also allowing for flexibility in what data is captured. For this reason, the MySQL team works to deliver the best of the SQL and Non-SQL worlds to our users, including native, NoSQL access to MySQL storage engines, with benchmarks showing 9x higher INSERT rates than using SQL, while also supporting online DDL.
At the same time, we do not sacrifice data integrity by trading away ACID compliance, and we do not trade away the ability to run complex SQL-based queries across the same data sets. This approach enables developers of new services to get the best out of MySQL database technologies.
Q4. Playful Play, a Mexico-based company, is using MySQL Cluster Carrier Grade Edition (CGE) to support three million subscribers on Facebook in Latin America. What are the technical challenges they are facing for such project? How do they solve them?
Tomas Ulin: As a start-up business, fast time to market at the lowest possible cost was their leading priority. As a result, they developed the first release of the game on the LAMP stack.
To meet both the scalability and availability requirements of the game, Playful Play initially deployed MySQL in a replicated, multi-master configuration.
As Playful Play’s game, La Vecidad de El Chavo, spread virally across Facebook, subscriptions rapidly exceeded one million users, leading Playful Play to consider how to best architect their gaming platforms for long-term growth.
The database is core to the game, responsible for managing:
• User profiles and avatars
• Gaming session data;
• In-app (application) purchases;
• Advertising and digital marketing event data.
In addition to growing user volumes, the El Chavo game also added new features that changed the profile of the database. Operations became more write-intensive, with INSERTs and UPDATEs accounting for up to 70 percent of the database load.
The game’s popularity also attracted advertisers, who demanded strict SLAs for both performance (predictable throughput with low latency) as well as uptime.
After their evaluation, PlayFul Play decided MySQL Cluster was best suited to meet their needs for scale and HA.
After their initial deployment, they engaged MySQL consulting services from Oracle to help optimize query performance for their environment, and started to use MySQL Cluster Manager to manage their installation, including automating the scaling of their infrastructure to support the growth from 30,000 new users every day. With subscriptions to MySQL Cluster CGE, which includes Oracle Premier Support and MySQL Cluster Manager in an integrated offering, Playful Play has access to qualified technical and consultative support which is also very important to them.
Playful Play currently supports more than four million subscribers with MySQL.
More details on their use of MySQL Cluster are here.
Q5. What are the current new commercial extensions for MySQL Enterprise Edition? How do commercial extensions differ from standard open source features of MySQL?
Tomas Ulin: MySQL Community Edition is available to all at no cost under the GPL. MySQL Enterprise Edition includes advanced features, management tools and technical support to help customers improve productivity and reduce the cost, risk and time to develop, deploy and manage MySQL applications. It also helps customers improve the performance, security and uptime of their MySQL-based applications.
This link is to a short demo that illustrates the added value MySQL Enterprise Edition offers.
Further details about all the commercial extensions can be found in this white paper. (Edit: You must be logged in to access this content.)
The newest additions to MySQL Enterprise Edition, released last fall during MySQL Connect, include:
* MySQL Enterprise Audit, to quickly and seamlessly add policy-based auditing compliance to new and existing applications.
* Additional MySQL Enterprise High Availability options, including Distributed Replicated Block Device (DRBD) and Oracle Solaris Clustering, increasing the range of certified and supported HA options for MySQL.
Q6. In September 2012, Oracle announced the first development milestone release of MySQL Cluster 7.3. What is new?
Tomas Ulin: The release comprises:
• Development Release 1: MySQL Cluster 7.3 with Foreign Keys. This has been one of the most requested enhancements to MySQL Cluster – enabling users to simplify their data models and application logic – while extending the range of use-cases.
• Early Access “Labs” Preview: MySQL Cluster Auto-Installer. Implemented with a standard HTML GUI and Python-based web server back-end, the Auto-Installer intelligently configures MySQL Cluster based on application requirements and available hardware resources. This makes it simple for DevOps teams to quickly configure and provision highly optimized MySQL Cluster deployments – whether on-premise or in the cloud.
Additional details can be found here.
Q7. John Busch, previously CTO of former Schooner Information Technology commented in a recent interview (1): “legacy MySQL does not scale well on a single node, which forces granular sharding and explicit application code changes to make them sharding-aware and results in low utilization of severs”. What is your take on this?
Tomas Ulin: Improving scalability on single nodes has been a significant area of development, for example MySQL 5.6 introduced a series of enhancements which have been further built upon in the server, optimizer and InnoDB storage engine. Benchmarks are showing close to linear scalability for systems with 48 cores / threads, with 230% higher performance than MySQL 5.5. Details are here.
Q8. Could you please explain in your opinion the trade-off between scaling out and scaling up? What does it mean in practice when using MySQL?
Tomas Ulin: On the scalability front, MySQL has come a long way in the last five years and MySQL 5.6 has made huge improvements here – depending on your workload, you may scale well up to 32 or 48 cores. While the old and proven techniques can work well as well: you may use master -slave(s) replication and split your load by having writes on the master only, while reads on slave(s), or using “sharding” to partition data across multiple computers. The question here is still the same: do you hit any bottlenecks on a single (big) server or not?… – and if yes, then you may start to think which kind of “distributed” solution is more appropriate for you. And it’s not only MySQL server related — similar problems and solutions are coming with all applications today targeting a high load activity.
Some things to consider…
- easier management
- easier to achieve consistency of your data
- you need to dimension your server up-front, or take into account the cost of throwing away your old hardware when you scale
- cost/performance is typically better on smaller servers
- in the end higher cost
- at some point you reach the limit, i.e. you can only scale-up so much
- you can start small with limited investment, and invest incrementally as you grow, reusing existing servers
- you can choose hardware with the optimal cost/performance
- more complicated management
- you need to manage data and consistency across multiple servers, typically by making the application/middle tier aware of the data distribution and server roles in your scale-out (choosing MySQL Cluster for scale-out does not incur this con, only if you choose a more traditional master-slave MySQL setup)
Q9. How can you obtain scalability and high-performance with Big Data and at the same time offer SQL joins?
Tomas Ulin: Based on estimates from leading Hadoop vendors, around 80 percent of their deployments are integrated with MySQL.
As discussed above, there has been significant development in NoSQL APIs to InnoDB and MySQL Cluster storage engines which allow high speed ingestion of high velocity data Key/Value data, but which also allows complex queries, including JOIN operations to run across that same data set using SQL.
Technologies like Apache Sqoop are commonly used to load data to and from MySQL and Hadoop, so many users will JOIN structured, relational data from MySQL with unstructured data such as clickstreams within Map/Reduce processes within Hadoop. We are also working on our Binlog API to enable real time CDC with Hadoop. More on the Binlog API is here.
Q10. Talking about scalability and performance what are the main differences if the database is stored on hard drives, SAN, flash memory (Flashcache)? What happens when data does not fit in DRAM?
Tomas Ulin: The answer depends on your workload.
The most important question is: “how long is your active data set remaining cached?” — if not long at all, then it means your activity will remain I/O-bound, and having faster storage here will help. But, if the “active data set” is small enough to be cached most of the time, the impact of the faster storage will not be as noticeable. MySQL 5.6 delivers many changes improving performance on heavy I/O-bound workloads.
Mr. Tomas Ulin has been working with the MySQL Database team since 2003. He is Vice President for the MySQL Engineering team, responsible for the development and maintenance of the MySQL related software products within Oracle, such as the MySQL Server, MySQL Cluster, MySQL Connectors, MySQL Workbench, MySQL Enterprise Backup, and MySQL Enterprise Monitor. Prior to working with MySQL his background was in the telecom industry, working for the Swedish telecom operator Telia and Telecom vendor Ericsson. He has a Masters degree in Computer Science and Applied Physics from Case Western Reserve University and a PhD in Computer Science from the Royal Institute of Technology.
ODBMS.org: Relational Databases, NewSQL, XML Databases, RDF Data Stores:
Blog Posts | Free Software | Articles and Presentations| Lecture Notes | Tutorials| Journals |
Follow us on Twitter: @odbmsorg
“Lufthansa is now able to aggregate and feed data into a management cockpit to analyze collected data for key decision-making purposes in the future. Users get instantly notified of transmission errors, enabling the company to detect patterns on large amounts of data at a rapid speed. There is also an automatic alarm messages sent out to IT product management, and partner airlines are informed of errors right away in the case of transmission errors between different IT systems for passenger data. Lufthansa is now able to comprehensively monitor one of its most important core processes in real-time for quality management: the handover of passenger data between different airlines” – James Dixon.
On the state of the market for Big Data Analytics I have Interviewed James Dixon, co-founder and Chief Geek / CTO, Pentaho Corporation.
Q1. What is In your opinion the expected realistic Market Demand for Big Data analytics?
James Dixon: Big. Until recently it has not been possible to perform analysis of sub-transactional and detailed operational data for a reasonable price-tag. Systems such as Hadoop and the NoSQL repositories such as MongoDB and Cassandra make it possible to economically store and process large amount of data. The first use of this data is often to answer operational and tactical questions. Shortly after that comes the desire to answer managerial and strategic questions, and this where Big Data Analytics comes in. I estimate that 90% of all Big Data repositories will have some form of reporting/visualization/analysis requirement applied to it.
Q2. Aren’t we too early with respect to the maturity of the Big Data Analytics technology and the market acceptance?
James Dixon: We are early, but not too early. There is significant market acceptance in certain domains already – financial services, SaaS application providers, and media companies to name a few. As these initial markets mature we will see common use cases emerge and public endorsements of these technologies, this will help to increase acceptance in other markets. We’ve seen a significant uptake in commercial deals over the last few quarters whereas 2011 was more tire-kicking and exploratory.
Q3. Pentaho has worked with Lufthansa to improve their passenger handling. Could you please tell us more about this? In particular what requirement and technical challenges did you have for this project? And how did you solve them?
James Dixon: Lufthansa needed a solution that would make the core processes of Inter Airline Through Check In (IATCI) accessible, measurable and available for real-time operational monitoring. They also wanted to deliver consolidated management reporting dashboards to inform decision making out of this information. This was implemented by Pentaho’s services organization with onsite training and consulting. Our Pentaho Business Analytics suite was used for the front-end for real-time data analysis and report generation. In the back-end, Pentaho Data Integration (aka Kettle) retrieves, transforms and loads the message data streams into the data warehouse on a continuous basis.
Q4. And what results did you obtain so far?
James Dixon: Lufthansa is now able to aggregate and feed data into a management cockpit to analyze collected data for key decision-making purposes in the future. Users get instantly notified of transmission errors, enabling the company to detect patterns on large amounts of data at a rapid speed. There is also an automatic alarm messages sent out to IT product management, and partner airlines are informed of errors right away in the case of transmission errors between different IT systems for passenger data. Lufthansa is now able to comprehensively monitor one of its most important core processes in real-time for quality management: the handover of passenger data between different airlines. With Pentaho, Lufthansa is now instantly aware if they are dealing with a single occurrence of an error or if there is a pattern. They can immediately take action in order to minimize the impact on their passengers.
Q5. What is special about Pentaho’s big data analytic platform? How does it differ with respect to other vendors?
James Dixon: We have an end-to-end offering that encompasses data integration/orchestration across Big Data and regular data stores/sources, data transformation, desktop and web-based reporting, slice-and-dice analysis tools, dash boarding, and predictive analytics. Very few vendors have the breadth of technology that we do, and those that do are mainly pushing hardware and services. We enable the creation of hybrid solutions that allow companies to use the most appropriate data storage technology for every part of their system – we don’t force you to load all your data into Hadoop, for example. From an architecture perspective our ability to run our data integration engine inside of MapReduce tasks on the data nodes is a unique capability. And we provide analytics directly on top of big data tech that gives users instant results via our schema-on-read approach – you don’t have to predefine ETL or Schemas or Data Marts – we do it on the fly.
Q6. What are the technical challenges in creating and viewing Analytics on the iPad?
James Dixon: The navigation concepts are different on mobile devices, so the overall user experience of the analysis software needs to be adapted for the iPad. Vendors need to be sensitive to the interaction techniques that the touch screens provide. We have changed the way that all of our end-user web-based interfaces work so that experience on the iPad is similar. It is possible to allow ad-hoc analysis and content authoring on the iPad, and Pentaho provides that with our recent V4.8 release.
Q7. Big Data and Mobile: what are the challenges and opportunities?
James Dixon: There are some use cases that are easy to identify. Report bursting to mobile and non-mobile devices is a technique that is easy to do today. Real-time analysis of Big Data combined with the alerting and notification capability of mobile devices is an interesting combination.
Q8. How are you supposed to view complex analytics with the limited display of a mobile phone?
James Dixon: Even with a desktop computer and a large monitor, analysis of Big Data requires lots of aggregation and/or lots of filtering. If you could display all the raw data from a Big Data repository, you would not be able to interpret it. As the display gets smaller the amount of aggregation and filtering has to go up, and the complexity has to come down. It is possible to do reasonably complex analysis on a tablet, but it is certainly a challenge on the smaller devices.
Q9. What are the main technical and business challenges that customers face when they want to use Cloud analytics deployments?
James Dixon: Moving large amounts of data around is a hurdle for some organizations. For this reason cloud analytics is not very appealing to companies with established data centers. However young companies that exclusively use hosted applications do not have their data on-premise. As these companies grow and mature we will see the market for cloud analytics increase.
Q10. Pentaho has announced in July this year a technical integration of their analytics platform with Cloudera. What is the technical and business meaning of this? What are results obtained so far?
James Dixon:We are working closely with Coudera on a technical and business level. For example we worked with Cloudera to test their new Impala database with Pentaho’s analytics, so that we could demo the integration on the day that Impala was announced. We also have joint marketing campaigns and sales field engagement, as customers of Cloudera find tremendous benefit in engaging with Pentaho and vice-versa. Our tech makes it much easier and 20x faster to get Hadoop productive so their customers gravitate to us naturally.
James Dixon, Founder and Chief Geek / CTO, Pentaho Corporation
As “Chief Geek” (CTO) at Pentaho, James Dixon is responsible for Pentaho’s architecture and technology roadmap. James has over 15 years of professional experience in software architecture, development and systems consulting. Prior to Pentaho, James held key technical roles at AppSource Corporation (acquired by Arbor Software which later merged into Hyperion Solutions) and Keyola (acquired by Lawson Software). Earlier in his career, James was a technology consultant working with large and small firms to deliver the benefits of innovative technology in real-world environments.
- Big Data: Challenges and Opportunities.
Roberto V. Zicari, October 5, 2012.
Abstract: In this presentation I review three current aspects related to Big Data:
1. The business perspective, 2. The Technology perspective, and 3. Big Data for social good.
Presentation (89 pages) | Intermediate| English | DOWNLOAD (PDF)| October 2012|
You can follow ODBMS.org on Twitter : @odbmsorg.
“There is only so much static data in the world as of today. The vast majority of new data, the data that is said to explode in volume over the next 5 years, is arriving from a high velocity source. It’s funny how obvious it is when you think about it. The only way to get Big Data in the future is to have it arrive in a high velocity rate ” – Scott Jarr.
One of the key technical challenges of Big Data is (Data) Velocity. On that, I have interviewed Scott Jarr, Co-founder and Chief Strategy Officer of VoltDB.
Q1. Marc Geall, past Head of European Technology Research at Deutsche Bank AG/London, writes about the “Big Data myth”, claiming that there is:
1) limited need of petabyte-scale data today,
2) very low proportion of databases in corporate deployment which requires more than tens of TB of data to be handled, and
3) lack of availability and high cost of highly skilled operators (often post-doctoral) to operate highly scalable NoSQL clusters.
What is your take on this?
Scott Jarr: Interestingly I agree with a lot of this for today. However, I also believe we are in the midst of a massive shift in business to what I call data-as-a-priority.
We are just beginning, but you can already see the signs. People are loathed to get rid of anything, sensors are capturing finer resolutions, and people want to make far more, data informed decisions.
I also believe that the value that corporate IT teams were able to extract from data with the advent of data warehouses really whet the appetite of what could be done with data. We are now seeing people ask questions like “why can’t I see this faster,” or “how do we use this incoming data to better serve customers,” or “how can we beat the other guys with our data.”
Data is becoming viewed as a corporate weapon. Add inbound data rates (velocity) combined with the desire to use data for better decisions and you have data sizes that will dwarf what is considered typical today. And almost no industry is excluded. The cost ceiling has collapsed.
Q2: What are the typical problems that are currently holding back many Big Data projects?
1) Spending too much time trying to figure out what solution to use for what problem. We were seeing this so often that we created a graphic and presentation that addresses this topic. We called it the Data Continuum.
2) Putting out fires that the current data environment is causing. Most infrastructures aren’t ready for the volume or velocity of data that is already starting to arrive at their doorsteps. They are spending a ton of time dealing with band-aids on small-data-infrastructure and unable to shake free to focus on the Big Data infrastructure that will be a longer-term fix.
3) Being able to clearly articulate the business value the company expects to achieve from a Big Data project has a way of slowing things down in a radical way.
4) Most of the successes in Big Data projects today are in situations where the company has done a very good job maintaining a reasonable scope to the project.
Q3: Why is it important to solve the Velocity problem when dealing with Big Data projects?
Scott Jarr: There is only so much static data in the world as of today. The vast majority of new data, the data that is said to explode in volume over the next 5 years, is arriving from a high velocity source. It’s funny how obvious it is when you think about it. The only way to get Big Data in the future is to have it arrive in a high velocity rate.
Companies are recognizing the business value they can get by acting on that data as it arrives rather than depositing it in a file to be batch processed at some later data. So much of the context that makes that data is lost when it not acted on quickly.
Q4: What exactly is Big Data Velocity? Is Big Data Velocity the same as stream computing?
Scott Jarr: We think of Big Data Velocity as data that is coming into the organization at a rate that can’t be managed in the traditional database. However, companies want to extract the most value they can from that data as it arrives. We see them doing three specific things:
1) Ingesting relentless feed(s) of data;
2) Making decisioning on each piece of data as it arrives; and
3) Using real-time analytics to derive immediate insights into what is happening with this velocity data.
Making the best possible decision each time data is touched is what velocity is all about. These decisions used to be called transactions in the OLTP world. They involve using other data stored in the database to make decision – approve a transaction, server the ad, authorize the access, etc. These decisions, and the real-time analytics that support them, all require the context of other data. In other words, the database used to perform these decisions must hold some amount of previously processed data – they must hold state. Streaming systems are good at a different set of problems.
Q5: Two other critical factors often mentioned for Big Data projects are: 1) Data discovery: How to find high-quality data from the Web? and 2) Data Veracity: How can we cope with uncertainty, imprecision, missing values, mis-statements or untruths? Any thoughts on these?
Scott Jarr: We have a number of customers who are using VoltDB in ways to improve data quality within their organization. We have one customer who is examining incoming financial events and looking for misses in sequence numbers to determine lost or miss-ordered information. Likewise, a popular use case is to filter out bad data as it comes in by looking at it in its high velocity state against a known set of bad or good characteristics. This keeps much of the bad data from ever entering the data pipeline.
Q6: Scalability has three aspects: data volume, hardware size, and concurrency. Scale and performance requirements for Big Data strain conventional databases. Which database technology is best to scale to petabytes?
Scott Jarr: VoltDB is focused on a very different problem, which is how to process that data prior to it landing in the long-term petabyte system. We see customers deploying VoltDB in front of both MPP OLAP and Hadoop, in roughly the same numbers. It really all depends on what the customer is ultimately trying to do with the data once it settles into its resting state in the petabyte store.
Q7: A/B testing, sessionization, bot detection, and pathing analysis all require powerful analytics on many petabytes of semi-structured Web data. Do you have some customers examples in this area?
Scott Jarr: Absolutely. Taken broadly, this is one of the most common uses of VoltDB. Micro-segmentation and on-the-fly ad content optimization are examples that we see regularly. The ability to design an ad, in real-time, based on five sets of audience meta-data can have a radical impact on performance.
Q8: When would you recommend to store Big Data in a traditional Data Warehouse and when in Hadoop?
Scott Jarr: My experience here is limited. As I said, our customers are using VoltDB in front of both types of stores to do decisioning and real-time analytics before the data moves into the long term store. Often, when the data is highly structured, it goes into a data warehouse and when it is less structured, it goes into Hadoop.
Q9: Instead of stand-alone products for ETL, BI/reporting and analytics wouldn’t it be better to have a seamless integration? In what ways can we open up a data processing platform to enable applications to get closer?
Scott Jarr: This is very much inline with our vision of the world. As Mike (Stonebraker , VoltDB founder) has stated for years, in high performance data systems, you need to have specialized databases. So we see the new world having far more data pipelines than stand alone databases. A data pipeline will have seamless integrations between velocity stores, warehouses, BI tools and exploratory analytics. Standards go a long way to making these integrations easier.
Q10: Anything you wish to add?
Scott Jarr.: Thank you Roberto. Very interesting discussion.
VoltDB Co-founder and Chief Strategy Officer Scott Jarr. Scott brings more than 20 years of experience building, launching and growing technology companies from inception to market leadership in highly competitive environments.
Prior to joining VoltDB, Scott was VP Product Management and Marketing at on-line backup SaaS leader LiveVault Corporation. While at LiveVault, Scott was key in growing the recurring revenue business to 2,000 customers strong, leading to an acquisition by Iron Mountain. Scott has also served as board member and advisor to other early-stage companies in the search, mobile, security, storage and virtualization markets. Scott has an undergraduate degree in mathematical programming from the University of Tampa and an MBA from the University of South Florida.
- Big Data: Challenges and Opportunities.
Roberto V. Zicari, October 5, 2012.
Abstract: In this presentation I review three current aspects related to Big Data:
1. The business perspective, 2. The Technology perspective, and 3. Big Data for social good.
Presentation (89 pages) | Intermediate| English | DOWNLOAD (PDF)| October 2012|
You can follow ODBMS.org on Twitter : @odbmsorg.
” We will observe at LEAST 1,000,000,000 celestial objects. If we launched today we would cope with difficulty – but we are on track to be ready by September when we actually launch. This is a game changer for astronomy thus very challenging for us, but we have done many large scale tests to gain confidence in our ability to process the complex and voluminous data arriving on ground and turn it into catalogues. I still feel the galaxy has plenty of scope to throw us an unexpected curve ball though and really challenge us in the data processing.” — William O`Mullane.
The Gaia mission is considered by the experts “the biggest data processing challenge to date in astronomy”. I recall here the Objectives and the Mission of the Gaia Project (source ESA Web site):
“To create the largest and most precise three dimensional chart of our Galaxy by providing unprecedented positional and radial velocity measurements for about one billion stars in our Galaxy and throughout the Local Group.”
“Gaia is an ambitious mission to chart a three-dimensional map of our Galaxy, the Milky Way, in the process revealing the composition, formation and evolution of the Galaxy. Gaia will provide unprecedented positional and radial velocity measurements with the accuracies needed to produce a stereoscopic and kinematic census of about one billion stars in our Galaxy and throughout the Local Group. This amounts to about 1 per cent of the Galactic stellar population. Combined with astrophysical information for each star, provided by on-board multi-colour photometry, these data will have the precision necessary to quantify the early formation, and subsequent dynamical, chemical and star formation evolution of the Milky Way Galaxy.
Additional scientific products include detection and orbital classification of tens of thousands of extra-solar planetary systems, a comprehensive survey of objects ranging from huge numbers of minor bodies in our Solar System, through galaxies in the nearby Universe, to some 500 000 distant quasars. It will also provide a number of stringent new tests of general relativity and cosmology.”
Last year in February, I have interviewed William O`Mullane, Science Operations Development Manager, at the European Space Agency, and Vik Nagjee, Product Manager, Core Technologies, at InterSystems Corporation, both deeply involved with the initial Proof-of-Concept of the data management part of the project.
A year later, I have asked William O`Mullane (European Space Agency), and Jose Ruperez (Intersystems Spain), some follow up questions.
Q1. The original goal of the Gaia mission was to “observe around 1,000,000,000 celestial objects”. Is this still true? Are you ready for that?
William O’Mullane: YES ! We will have a Ground Segment Readiness Review next Spring and a Flight Acceptance Review before summer. We will observe at LEAST 1,000,000,000 celestial objects. If we launched today we would cope with difficulty – but we are on track to be ready by September when we actually launch. This is a game changer for astronomy thus very challenging for us, but we have done many large scale tests to gain confidence in our ability to process the complex and voluminous data arriving on ground and turn it into catalogues. I still feel the galaxy has plenty of scope to throw us an unexpected curve ball though and really challenge us in the data processing.
Q2. The plan was to launch the Gaia satellite in early 2013. Is this plan confirmed?
William O’Mullane: Currently September 2013 in Q1 is the official launch date.
Q3. Did the data requirements for the project change in the last year? If yes, how?
William O’Mullane: Downlink rate has not changed so we know how much comes into the System still only about 100TB over 5 years. Data processing volumes depend on how many intermediate steps we keep in different locations. Not much change there since last year.
Q4. The sheer volume of data that is expected to be captured by the Gaia satellite poses a technical challenge. What work has been done in the last year to prepare for such a challenge? What did you learn from the Proof-of-Concept of the data management part of this project?
William O’Mullane: I suppose we learned the same lessons as other projects. We have multiple processing centres with different needs met by different systems. We did not try for a single unified approach across these centers.
The CNES have gone completely to Hadoop for their processing. At ESAC we are going to InterSystems Caché. Last year only AGIS was on Caché – now the main daily processing chain is in Caché also [Edit: see also Q.9 ]. There was a significant boost in performance here but it must be said some of this was probably internal to the system, in moving it we looked at some bottlenecks more closely.
Jose Ruperez: We are very pleased to know that last year was only AGIS and now they have several other databases in Caché.
William O’Mullane: The second operations rehearsal is just drawing to a close. This was run completely on Caché (the first rehearsal used Oracle). There were of course some minor problems (also with our software) but in general from Caché perspective it went well.
Q5. Could you please give us some numbers related to performance? Could you also tells us what bottlenecks did you look at, and how did you avoid them?
William O’Mullane: Would take me time to dig out numbers .. we got factor 10 in some places with combination of better queries and removing some code bottle necks. We seem to regularly see factor 10 on “non optimized” systems.
Q6. Is it technically possible to interchange data between Hadoop and Caché ? Does it make sense for the project?
Jose Ruperez: The raw data downloaded from the satellite link every day can be loaded in any database in general. ESAC has chosen InterSystems Caché for performance reasons, but not only. William also explains how cost-effectiveness as well as the support from InterSystems were key. Other centers can try and use other products.
This is a valid point – support is a major reason for our using Caché. InterSystems work with us very well and respond to needs quickly. InterSystems certainly have a very developer oriented culture which matches our team well.
Hadoop is one thing HDFS is another .. but of course they go together. In many ways our DataTrain Whiteboard do “map reduce” with some improvements for our specific problem. There are Hadoop database interfaces so it could work with Caché.
Q7. Could you tell us a bit more what did you learn so far with this project? In particular, what is the implication for Caché, now that also the the main daily processing chain is stored in Caché?
Jose Ruperez: A relevant number, regarding InterSystems Caché performance, is to be able to insert over 100,000 records per second sustained over several days. This also means that InterSystems Caché, in turn, has to write hundreds of MegaBytes per second to disk. To me, this is still mind-boggling.
And done with fairly standard NetApp Storage. Caché and NetApp engineers sat together here at ESAC to align the configuration of both systems to get the max IO for Java through Caché to NetApp. There were several low level page size settings etc. which were modified for this.
Q8. What else still need to be done?
William O’Mullane: Well we have most of the parts but it is not a well oiled machine yet. We need more robustness and a little more automation across the board.
Q9. Your high level architecture a year ago consisted of two databases, a so called Main Database and an AGIS Database.
The Main Database was supposed to hold all data from Gaia and the products of processing. (This was expected to grow from a few TBs to few hundreds of TBs during the mission). AGIS was only required a subset of this data for analytics purpose. Could you tell us how the architecture has evolved in the last year?
William O’Mullane: This remains precisely the same.
Q10. For the AGIS database, were you able to generate realistic data and load on the system?
William O’Mullane: We have run large scale AGIS tests with 50,000,000 sources or about 4,500,000,000 simulated observations. This worked rather nicely and well within requirements. We confirmed going from 2 to 10 to 50 million sources that the problem scales as expected. The final (end of mission 2018) requirement is for 100,000,000 sources, so for now we are quite confident with the load characteristics. The simulation had a realistic source distribution in magnitude and coordinates (i.e. real sky like inhomogeneities are seen).
Q11. What results did you obtain in tuning and configuring the AGIS system in order to meet the strict insert requirements, while still optimizing sufficiently for down-stream querying of the data?
William O’Mullane: We still have bottlenecks in the update servers but the 50 million test still ran inside one month on a small in house cluster. So the 100 million in 3 months (system requirement) will be easily met especially with new hardware.
Q12. What are the next steps planned for the Gaia mission and what are the main technical challenges ahead?
William O’Mullane: AGIS is the critical piece of Gaia software for astrometry but before that the daily data treatment must be run. This so called Initial Data Treatment (IDT) is our main focus right now. It must be robust and smoothly operating for the mission and able to cope with non nominal situations occurring in commissioning the instrument. So some months of consolidation, bug fixing and operational rehearsals for us. The future challenge I expect not to be technical but rather when we see the real data and it is not exactly as we expect/hope it will be.
I may be pleasantly surprised of course. Ask me next year …
William O`Mullane, Science Operations Development Manager, European Space Agency.
William O’Mullane has a PhD in Physics and a background in Computer Science and has worked on space science projects since 1996 when he assisted with the production of the Hipparcos CDROMS. During this period he was also involved with the Planck and Integral science ground segments as well as contemplating the Gaia data processing problem. From 2000-2005 Wil worked on developing the US National Virtual Observatory (NVO) and on the Sloan Digital Sky Survey (SDSS) in Baltimore, USA. In August 2005 he rejoined the European Space Agency as Gaia Science Operations Development Manager to lead the ESAC development effort for the Gaia Data Processing and Analysis Consortium.
José Rupérez, Senior Engineer at InterSystems.
He has been providing technical advise to customers and partners in Spain and Portugal for the last 10 years. In particular, he has been working with the European Space Agency since December 2008. Before InterSystems, José developed his career at eSkye Solutions and A.T. Kearney in the United States, always in Software. He started his career working for Alcatel in 1994 as a Software Engineer. José holds a Bachelor of Science in Physics from Universidad Complutense (Madrid, Spain) and a Master of Science in Computer Science from Ball State University (Indiana, USA). He has also attended courses at the MIT Sloan School of Business.
You can follow ODBMS.org on Twitter : @odbmsorg.