On MySQL HeatWave. Q&A with Frederic Descamps
Q1. What is exactly MySQL HeatWave and how does it compared to the “vanilla” MySQL?
MySQL HeatWave is the name of the MySQL Managed Service (DBaaS) by Oracle. Of course it’s available in OCI but also in AWS and Azure.
It’s the MySQL Team itself, operating and supporting it.
Compared to the MySQL Community Edition, the MySQL version running on HeatWave is the Enterprise Edition. But it comes with other linked products like HeatWave Cluster, that accelerates Analytics (OLAP) queries, certain OLTP can also benefit from it. HeatWave provides also several advisors where ML shines. For example the system can advise to create or remove some indexes, …
Another component of the solution is LakeHouse where the data is stored in Object Storage and available directly in MySQL. Recently we also have included GenAI in our cloud solution.
Q2. Please tell us how to run analytics queries (aka OLAP) using the HeatWave cluster?
Do you have any benchmark numbers on how fast are these queries answered?
To use it, you only need few steps. Of course you need to enable the cluster. You can enable it on demand and pay for what you need only when you need it. Once the cluster is available, with some few click in the interface or directly in MySQL Shell for Visual Studio Code, you can load the data you want into the cluster. As I replied in the previous question, you can ask advisor the best way to load your data and that process will return a simple SQL statement to run.
The nice thing with the solution is once loaded, the data keeps to be updated in nearly realtime. If you know MySQL, this is done consumming the binlogs like any other replication solution.
Once loaded, for the developper nothing else needs to be changed. If any type of query is sent to MySQL, via the MySQL protocol using the usual connection, it’s the MySQL optimizer that will decide if the query will benefit from being sent to the cluster or not. And once sent, you can imagine it is much faster! We have several benchmarks published on our website, with the source available on GitHub. Of course it all depends on the amount of data or type of queries but take a look at https://www.oracle.com/heatwave/performance-benchmarks/
Q3. But you can also run such queries on files using LakeHouse. What is LakeHouse and what is the difference with respect to using a HeatWave cluster?
So HeatWave Lakehouse processes data in a variety of file formats, including CSV, Parquet, Avro, JSON, and exports from other databases that you store into an Object Storage. Of course you can combine this data with your transactional data in MySQL databases directly too.
Lakehouse is an addition of the HeatWave Cluster.
Q4. When we talk about analytics, we also think about data visualization solution. How important is data visualization?
Data visualization is a bit out of my expertize domain, but what could you do will such amount of data that all systems are collecting without having tools to vizualize it? I think this is more and more important and this is again a domain where AI is helping a lot because finding the best way to visualize data is not a straightforward process… I tried several different solutions, they are powerful but I wouldn’t call them easy.
Q5. Why is it important to establish a connection between Oracle Analytics Cloud (OAC) and a MySQL HeatWave DB instance within Oracle Cloud Infrastructure (OCI)?
For security reasons, the MySQL managed service is OCI doesn’t allow any public access. Running the database open to the Internet is not a very safe architecture. I know a lot of people that first test their database in cloud services where it’s possible to let it open publicly and finally never fix that… and later could pay the price of that weak security choice. To avoid all those kind of problems we just don’t allow it. That’s the reason we need to add an access between OAC and the VCN where the DB System is located.
Q6. Where can we find a step by step example of such procedure?
I’ve wrote a blog explain this.
Q7. What are the pros and cons of this solution?
I would say that the only cons could eventually be that the process is not straightforward out-of-the-box as you indeed need to create an OAC Private Access Channel. Other than that, as it’s more secure and you don’t breach your data out of the Internet, I can only see advantages. And then connected, OAC allows you that a large panel of vizualization methods for your data.
Qx. Anything else you wish to add?
We recently released the version 9.0 of MySQL that includes the VECTOR DATATYPE and in combination to HeatWave provides you very nice and robust solution for those planning to write GenAI applications. And for those willing to test MySQL HeatWave, we now offer an always free MySQL HeatWave instance. People can use the MySQL DBaaS and the HeatWave accelerator for free and in combination with the always free tier compute instance, any solution like WordPress can be easily deploy on OCI for free!
Resources
lefred’s blog: tribulations of a MySQL Evangelist
……………………………………………………
Frederic Descamps, MySQL Community Manager
lefred has been consulting OpenSource and MySQL for almost 20 years. After graduating in Management Information Technology, Frédéric started his career as a developer for an ERP under HPUX. He will then opt for a career in the world of open-source by joining one of the first Belgian start-up dedicated 100% to free projects around GNU/Linux. It is in 2011 that lefred joined Percona, one of the leading MySQL-based specialists. He decided to join the MySQL Community Team in 2016 as a MySQL Community Manager for EMEA & APAC. Frédéric is also a regular speaker of OpenSource Conferences. His blog mostly dedicated to MySQL is https://lefred.be
Fred is also the devoted father of three adorable daughters: Wilhelmine, Héloïse & Barbara.
Sponsored by MySQL/Oracle.