Oracle Claims Leadership In Cloud Data Warehouse Space Over Snowflake With Its New MySQL Heatwave Service

By Patrick Moorhead - September 2, 2021

Oracle announced its next-generation MySQL HeatWave offering today, focusing on machine learning (ML)-based automation and scalability that the company claims will drive greater performance and price-performance leadership in the market. As part of today's news, a raft of new publicly accessible and repeatable benchmarks became available, demonstrating Oracle's claim that MySQL HeatWave is an incredible seven times faster than Snowflake at 1/5 the cost and equally faster than other cloud services. These are bold claims. Do they hold water?

In the next few paragraphs, I take a deeper look into these new price/performance advantage in the MySQL space and dig into what’s new in MySQL Heatwave.

Setting the stage – what is HeatWave?

Before answering this question, maybe the best approach is to describe the problem statement. MySQL is deployed everywhere. Perhaps not *everywhere*, but close to it. It is in companies of all sizes, across all industries, and in every region of the world. Further, the data that has been stored in MySQL for the last 25+ years is vast and largely untapped. The database was designed for departmental-level use, running online transactional processing (OLTP) applications. So, running deep analytics or even complex OLTP queries could turn frustrating for many organizations. This is especially true when one considers that some enterprises can have thousands of MySQL instances running across departments, each with valuable data. Yeah, one could use an extract, transform, load (ETL) tool to move MySQL data into an analytics engine, but by the time this operation was completed, the data could be stale and the results outdated.

Oracle's solution to this was to design and build a dedicated analytics engine that could take MySQL data from databases across the enterprise and store it in memory in real-time. The result of this effort is known as MySQL HeatWave, a cloud-born service that enables an organization to extract the full potential from its MySQL environment.

Upon first hearing about HeatWave and Oracle's claims, I was skeptical. However, the numbers bear out, and Oracle publishes its testing scripts, configurations, and data to GitHub, which customers can review, replicate, and challenge. You can find deeper analysis on HeatWave here and gain access to Oracle's HeatWave GitHub postings here.

HeatWave's newest innovation –

Customers I've spoken with attest to the performance and cost savings that Oracle claimed. And this was a big step forward for the MySQL market. Arguably the biggest ever.

HeatWave's latest release seems to focus on what the company anticipated customers would want. Automation. Automation across the HeatWave service lifecycle. From set up to optimizing to performance and resiliency. MySQL customers looking to utilize the analytics capabilities in the cloud, be it from Oracle, Snowflake, Azure's Synapse, or GCP's BigQuery, can attest to the complexity of lighting up a fully optimized environment.

This goes beyond making sure that the right calls are made from on-prem to off-prem. This is about ensuring the appropriate level of normalization, assigning keys, and establishing and enforcing proper key relationships. And Oracle seems to have an answer with its latest release of HeatWave.

It's all about Autopilot

The heart of HeatWave's new release is MySQL Autopilot, a machine learning (ML)-powered optimization engine that delivers nine new automation services spanning the data management lifecycle.

MySQL Autopilot – ML driven MySQL optimization ORACLE

MySQL Autopilot uses a combination of deep statistics analysis and distinct ML models to optimize MySQL on a per-instance basis. As you can see in the graphic below, Autopilot extrapolates statistics from the MySQL environment. These statistics are used to feed these ML models to help feed and tune MySQL. Even if Autopilot resulted in no additional performance gains, I believe data consumers would find these capabilities invaluable.

Every database manager or the person charged with database performance knows that database tuning is a full-time job with equal parts technical, art form, and clairvoyance.

MySQL Autopilot takes this off the plate of database managers, allowing them to focus more on the strategic imperatives of the business.

The MySQL Autopilot Architecture ORACLE

There's more

While MySQL HeatWave receives a lot of attention for its capabilities, there's a bit more to this new MySQL HeatWave service release. Oracle has not only spent much time tuning performance but also enabling larger data sizes, increasing the number of nodes, accelerating more constructs, further increasing scalability, and securing data "in-flight."

HeatWave uses commodity servers powered by AMD EPYC processors, combined with an object store, where data is partitioned (and stored) in encrypted in-memory format. By employing a scale-out data management layer, users should see accelerated operations across their MySQL environments while reducing costs (compared to other cloud services).

While many more words can be spent on what Oracle has done with HeatWave, I can quickly sum up my thoughts by saying it's clear that Oracle is thinking about its customers, as every update seems to consider how a user would interact with, use, and manage its HeatWave environment.

Is HeatWave *that* much better performing?

This is always the "yeah, but…" section of any analysis done. Perhaps this section could have been more aptly titled, "did all of these neat innovations matter?" And as a former product marketeer (and IT executive), I view all performance claims with a bit of skepticism. But as I also said at the outset, the MySQL team took the unusual step of publishing its benchmarking results, along with scripts and environmentals, so that anybody can reproduce them. The results are impressive. In TPC-H, a business decision benchmark, Oracle demonstrates a price-performance advantage of a staggering 35x better price/performance ratio v Snowflake. This is important as this benchmark mimics a lot of the ad-hoc querying and operations a business would perform on a day-to-day basis.

Comparing HeatWave against Snowflake on TPC-H MOOR INSIGHTS & STRATEGY AND ORACLE

The benefit that HeatWave can bring to an organization looking to run analytics in its MySQL environment is impressive. To see even more comparisons, visit Oracle’s HeatWave page

Final thoughts

I believe "the cloud" has done an excellent job of driving home the economics of technology. IT organizations are becoming smarter about understanding the ROI on investments and ensuring that it is getting fair value. With this said, it's hard to look at what Oracle is doing with HeatWave and not be impressed. In many ways, HeatWave is the very essence of what the cloud is about – commodity infrastructure architected in a scale-out fashion, enabling the best performance and price-performance for a given workload (MySQL). 

Oracle is showing its chops in two areas – building a world-class cloud with Oracle Cloud Infrastructure (OCI) and continuously demonstrating its deep technical expertise in databases with this latest release of HeatWave. With Autonomous Database at the high-end, tier one segment of the enterprise market and HeatWave providing extremes of performance and cost differentiation for open source developers and cloud-born, emerging companies, Oracle provides customers with two well-differentiated and compelling options. The company seems to be hitting the competition with innovation from both angles. 

Customer-centric design isn’t always a good descriptor of technology companies in the IT solutions space. I believe Oracle demonstrated this very concept with HeatWave. The issue of quickly and easily extracting data from one’s MySQL environment has been an issue since MySQL gained popularity. Oracle answered. And it’s clear the company has been paying attention. 

I'm excited to see what comes next. 

Note: Moor Insights & Strategy writers and editors may have contributed to this article. 

+ posts

Patrick founded the firm based on his real-world world technology experiences with the understanding of what he wasn’t getting from analysts and consultants. Ten years later, Patrick is ranked #1 among technology industry analysts in terms of “power” (ARInsights)  in “press citations” (Apollo Research). Moorhead is a contributor at Forbes and frequently appears on CNBC. He is a broad-based analyst covering a wide variety of topics including the cloud, enterprise SaaS, collaboration, client computing, and semiconductors. He has 30 years of experience including 15 years of executive experience at high tech companies (NCR, AT&T, Compaq, now HP, and AMD) leading strategy, product management, product marketing, and corporate marketing, including three industry board appointments.