• Paul Burgess

Database Trending - an interesting perspective

Updated: Apr 24

Human nature is to measure - good versus bad, better versus worse, consider the references to Bo Derek's movie '10' or David Letterman's 'Top 10 lists'. Those social references are just a reflection in how we make technology decisions today - especially databases.. Rather it is benchmarking with TPC results, where a vendor falls in Gartner Magic Quadrant charts or the 'score cards' a team uses to put measures to Request for Proposals (RFP)s. The number of clients that would hire my employers over the years to be that 'outside voice' were usually to negate the political fallout if something ever went wrong - blame 'the consultant'. In the end, no know wants to risk their career.

A measure that has come to light worth calling out when it comes to data management is from the DB-Engines website. It covers about anyone you can imagine from a detailed vendor perspective - even IMS is still listed and that was released over 50 years ago. What the listing also gives us is a comparison to appreciate both point-in-time comparisons and a technologies rise or drop over time. If we consider the root problem with benchmarking, vendor reviews and RFPs they are a snapshot in time when they are published. DB-Engines adds to this change-over-time.


They even hand out awards titled 'Database of the Year' based on their own analysis (versus other publications who are pay-to-win types). For instance the last 5 years are:

Vendor

Year

Snowflake

2021

Postgres

2020

MySQL

2019

Postgres

2018

Postgres

2017

This author also likes to add some observations and viewpoints:s.

Vendor

Comments

Amazon

They have such a large catalog of technology - and databases - they cannot do any of it well. Consider the stagnation of so many of their repositories.

Microsoft

Their flagship SQLServer is in the top 5, but their 2012 version goes end of life this summer and Microsoft's preferred go-forward is to move those workloads into their Azure cloud. Suspect the status for SQLServer to go down but because not everyone is on Azure, do not predict the Azure cloud options to rise at the same pace.

Oracle

From the moment I met Larry Ellison in 1995, there has been a stigma and distrust for the entire company and the man. Fast forward to 2022 and even though they are still #1, suspect their reckoning as they only want on-premise Oracle licenses to the tier 1 workloads in the cloud are on the Oracle cloud. They purchased Peoplesoft, Siebel and other technologies to drive database consumption. But those same technologies are going away as enterprises move to the cloud. Add let us now forget their ownership of MySQL and the eroding loss in confidence for that technology at the hands of Larry. Refer to this article for more on this front.

Snowflake

Are the most compelling here as DB-Engines voted them the 'Database of the Year' in 2021. From a measurement perspective, their meteoric rise in the ranks over the last few years is more impressive. April 2020 their score was 2.4 (#117th on the list). In April 2021 they are 89.45 (and #14th on the list). Referring to the diagram above, they moved from #29 to #14 in the 12 months. I recommend organizations find out for themselves why this is and report back to me as this author was taken away much to be impressed with Snowflake.

Data Bricks

Of the 391 repositories listed, they are non-existent. They should stay in their core competency of Spark.

Teradata

The move to the cloud is not treating this legacy on-premise appliance stalwart well. They are laying off employees (during a labor shortage), they maintain a premium price for their cloud offerings called Vantage and earlier this year they initiated a $250 million stock buy back. These all add up to lost customer revenues and confidence.

Google

They are mentioned last because at least Azure and Amazon attempt to have a partner program to round out their platform consumption. The experiences and observations of Google are of a lack of anyone else to make money on their platform. They have a 'playing in the GCP sandbox' problem.

There are so many others not referenced above because of the lack of relevance in the modern technology world or their use is more niche (NoSQL, time series, etc). Consider:

  • IBM sold off Informix engineering and support to HCL. They acquired, killed and brought back Netezza. Add to it the IBM Cloud Pak for Data which is a way to bring loads of old tech together and make it look new. Reminds us of the collapsed mortgage bond market of 2008.

  • Postgres is the only true open-source technology showing forward momentum but because AWS, Azure and GCP all have a multitude of offerings. The other players like EnterpriseDB and Crunchy (who represent a majority of the contributions to the project) are all challenged with market differentiation better versus the cloud offerings.

  • SAP's ownership of Sybase (now called Adaptive Server) will continue to drop with the announced end-of-life to ASE by 2025.

  • It looks like Single Store (the rebranded MEMSql technology) has received investments from IBM and now being resold by IBM.

This summer should be interesting time to watch what happens around IPOs, acquisitions and how the old-try-to-look-new again.

10 views0 comments