Cisco & ParStream: A closer look at THE platform for IoT analytics in Real-Time
by Ramesh Chitor, Senior Business Development Manager, Cisco Systems. -December 2014.
“ParStream promises to be a key Cog for the IoT Wheel”
It has been a long time since the PC dominated the world of connected devices. And while smartphones and tablets have stolen the limelight in recent years they are also now taking a back seat as a new wave of connectivity takes hold. With the Internet of Things and Machine -2-M now in the vanguard, smart homes and wearables are also beginning to make their presence across devices felt. According to recent industry reports, there will be nearly three times as many connected devices in use by 2020 as in 2014, equivalent to more than four internet devices for every person on the planet.
• ParStream is the first platform, that provides Real-Time Analytics for the Internet of Things
• Based on ParStream, Cisco can improve existing products and enable new types of services in several areas
• Building a fast, reliable and highly scalable analytics infrastructure for fog-computing
• Implementing new types of real time network monitoring
• Providing analytical services in the cloud and software offerings on server clusters
Analytics in the Internet of Everything
IoT is outgrowing the Internet, as we know it by a factor of 6-8. It will consist of hundreds of billions of sensors and generate Zettabytes of Data.
There has been a strong focus on collecting data from the many sensors and smart devices in IoT, however, the ability to analyze these large amounts of data in real-time is the key success factor for getting the maximum value from IoT. It’s all about getting actionable insights from your data…which means it has to be analyzed quickly enough to drive action.
ParStream’s unique offering
ParStream provides a Big Data Analytics Platform for IoT, built around a unique, patented database technology.
ParStream DB is a columnar In-Memory database with a lockless, highly parallel and fault tolerant architecture. Based on patented algorithms for indexing and compressing data, ParStream is able to outperform any other technology by orders of magnitude (400-800 times fast than relational DB, >30.000 times faster than Hadoop/Hive).
There are six key requirements needed for IoT Analytics that ParStream provides in a unprecedented way
1.) Scalability; ParStream handles massive amounts of data, especially critical for IoT use-cases, where data will to continue to grow exponentially.
2.) Velocity of data ingestion; ParStream is able to load data with more than 2 million rows per second, on a single server, without affecting query performance.
3.) Speed in performing queries for immediate insights; ParStream is capable to execute instant ad hoc queries, which is critical for many IoT use-cases.
4.) Flexibility and portability of the analytics platform; ParStream is running on standard Linux operating systems, on single servers in cluster and in the cloud.
5.) Distributed Query processing; Due to the amount of data generated, data processing and analytics has to take place at the location of its origin, that is what Cisco calls Fog computing. ParStream is able to push down query execution to the place where data is generated and execute analytics in a highly distributed fashion.
6.) Efficiency of the footprint; ParStream is totally build by ParStream C++ and has a small footprint. Therefore it even can run in embedded system and on IOX.
ParStream is a proven, highly reliable system, running 24×7 in production environments at customer sites since years.
Real customer use cases include Gas Turbine Monitoring, Wind Turbine Park analysis, and Mobile Network monitoring and chip test analytics.
The relationship between ParStream and Cisco is growing fast. Several activities have been started and planned.
• ParStream IoT Analytics Platform was officially launched at the Cisco IoT World conference in Chicago, getting very positive response from Media and Analysts
• ParStream is part of the current Cisco Entrepreneur in Residence Program (EIR)
• Participation at the Cisco Systems IoT SEVT Partner Pavilion Event in Dec’2014 with key activity threads in progress..
• A PoC of ParStream in the Cisco IOE Vertical Solutions group in works.
• A demo setup for ParStream running on Cisco’s UCS server will start beginning of 2015 (together with Channel and SI partners)
With all the focus now shifting to specific customer use cases, 2015 promises a super exciting year ahead for the IoT community.
Sponsored by CISCO