MySQL HeatWave is a cloud database service that has enabled Oracle to create differentiation in the open-source MySQL market. With the latest announcement that MySQL HeatWave is available natively on AWS, Oracle looks to accelerate that separation from other cloud database providers. What exactly is HeatWave, and what did Oracle announce? I'll try to answer all that and more in the next few paragraphs.
MySQL HeatWave – an overview
Before getting into the details of the announcement, it's a good idea to spend a few words on what MySQL Heatwave is. And to do this, let's start by outlining the data challenge in virtually every enterprise.
Data is being generated everywhere in the enterprise, from applications, business units, user interactions, devices, partners and more. There is no shortage of databases and analytics tools to store and sort through the data. And eventually, enterprises end up with several of them, in some cases encouraged by cloud database providers who promote specialized databases for each use case.
Getting an overall view of the business and making decisions based on that view becomes a problem; enterprises struggle to bring together disjointed data in silos and different databases. To succeed, businesses must turn all this data into intelligence and actionable insights so they can make decisions in real-time. However, the extract, transform, and load (ETL) processes associated with merging data from multiple sources and formats is time-consuming, error-riddled and leaves security gaps.
Enter MySQL HeatWave. First introduced in December of 2020, HeatWave delivers online transaction processing (OLTP) and online analytics processing (OLAP) in one service in real-time with high levels of performance, security and reliability. Oracle’s positioning of HeatWave is as follows: 100% in-memory, all the time, any time – no excuses. For customers relying on accurate and timely intelligence, this is critical.
Since the initial release of HeatWave, the MySQL engineering team has continued to add increasing levels of functionality to HeatWave with a regular cadence of service upgrades, including Autopilot (machine learning–based automation for many data management functions), scale, security, elasticity, integrated machine learning with HeatWave ML and many other capabilities to increase performance and reliability.
The above may sound like hyperbole, but when one considers the value of data and the importance of what that data becomes – actionable intelligence – one can understand the importance of a database engine that can help a business find the proverbial needle in a haystack faster and use this intelligence more meaningfully.
As one would expect from a company as bold as Oracle, it has lined up benchmark after benchmark to show how much enterprises can benefit from HeatWave—not only in terms of raw performance but also price performance. And while the company claims incredible advantages over the competition, its comparison to Snowflake (Medium size shape) is worth mentioning – 16x better raw performance and 10x better price performance when running the TPC-H benchmark against a 4TB sample. While these numbers are startling, and as an analyst (and ex-IT pro), I always look at benchmarks with skepticism, Oracle has taken the unprecedented step of publishing the benchmark parameters and scripts on GitHub. So, if you share my benchmark cynicism, click hereand run them for yourself.
Reality check – we live in a multi-cloud world
As MySQL HeatWave gains prominence, more customers want to enjoy its benefits natively. In Oracle’s ideal world, every customer would move their apps and data to Oracle Cloud Infrastructure (OCI) to take advantage of HeatWave. And in fact, many have, contributing partly to the growth of OCI reported every quarter.
However, for as many customers that have taken advantage of MySQL HeatWave on OCI, there are many more with their apps and data sitting in AWS, Azure or some other cloud. These businesses can't or won't migrate to OCI for any number of reasons – regulatory, data sovereignty, data egress costs, apps tightly knit with cloud services and ecosystem or simply organizational inertia. This is the reality of living in a multi-cloud world.
This dynamic has led to a pragmatic and strategic move by Oracle, and that is to offer MySQL HeatWave on AWS. HeatWave is available natively on AWS, meaning that the data plane, control plane and the console run on AWS and have been optimized specifically for AWS. This results in similar levels of performance and price performance that Oracle has shown on OCI. Furthermore, AWS customers get the benefits of automation provided by MySQL Autopilot, and ETL-less processing by a single database for OLTP, OLAP and machine learning.
The above diagram shows how users interface with MySQL HeatWave on AWS and how HeatWave interacts with customers' apps and data that also run on AWS. As you can see, HeatWave runs wholly in AWS, including the control plane, data plane and is accessed by the HeatWave interactive console. In addition, HeatWave has been optimized for the AWS cloud and has newly added capabilities, including Auto Thread Pooling to prevent resource contention, auto shape prediction, security features like data masking and performance monitoring
While performance and price performance are obvious benefits of running HeatWave natively in AWS, I am equally impressed by the fact that HeatWave is a fully managed service just as on OCI. For any database administrator tasked with provisioning, tuning and maintaining an optimally performing (and performant) database environment, HeatWave with MySQL Autopilot on AWS is a gift from the database gods. The MySQL Autopilot automated management capabilities – which are workload-dependent and powered by machine learning models – can save hours upon hours and thousands of dollars of manual labor every week.
So, AWS is one cloud. But what about Azure? What about other clouds? Oracle has stated that MySQL HeatWave will be available for Azure in the near future, just as Autonomous Database and Exadata Database Service are available for Azure today, and I suspect that the company will support other clouds based on customer demand.
One of the changes I’ve seen in Oracle as a company that has impressed me is this understanding that the world around Oracle is changing. And maintaining (and increasing) relevance in this new world requires a different approach to how products and services are offered and consumed. The MySQL HeatWave on AWS announcementis an excellent example of how Oracle has evolved, as is the recent joint announcementwith Microsoft for Oracle Database Service for Microsoft Azure.
What this means for enterprise customers
Enterprise customers now have a choice of where to take advantage of MySQL HeatWave’s benefits. Oracle customers who want to use MySQL HeatWave but are running into resistance from teams that have built gravity around AWS will be thrilled. But really, this goes beyond Oracle customers. Any enterprise looking to simplify their data management and take advantage of running OLTP, real-time analytics and machine learning – plus automated management tasks in a single database – should look at what Oracle is doing with MySQL HeatWave on AWS.
Just in case my not-so-subtle messages were missed–Oracle is a changed company and MySQL HeatWave is proof of this evolution. While the company made its mark as the leader in the enterprise data management space, it has shaped its offerings to better support customers in their new world of data – structured, unstructured, transaction processing, analytics, machine learning, IOT and more – in a choice of on-premises, hybrid, public cloud and multi-cloud deployments. Deep database engineering efforts, led by Nipun Agarwal, have driven this evolution at Oracle. The innovations we see in HeatWave are the result of years of experience solving data management challenges in the enterprise, combined with an innovative mindset and an understanding of where the market is going.
I’ll be tracking the continued progress of MySQL HeatWave and its adoption across the enterprise and the clouds. Stay tuned.
Note: Moor Insights & Strategy writers and editors may have contributed to this article.