Opinion Last week, I committed a crime. No, I did not defraud the IRS/HMRC or steal some Snyder’s Pretzels/Cheesy Wotsits. It was much worse. A keen reader pointed out that I had miscounted the number of versions of Db2, IBM’s relational database.
Although I had listed the versions of IBM’s 40-year-old database running on Linux or Unix or Windows (LUW) and its System Z mainframe, I neglected to mention the version running on the mid-range server once known as AS/400, now – after several intermediary name changes – called iSeries or IBM i.
Shamed though I was by my error, I could at least find comfort in the knowledge that I had not – like IBM to all outward appearances – forgotten about Db2 altogether.
For more than four years, I have been asking, enticing, cajoling, coaxing, prompting, tempting, persuading, encouraging, inciting, and seducing IBM to talk to The Register about Db2. I’ve tried the front door, the back door, the side door, the trap door, the doors of perception, the garden gate, and the sunroof. It’s all been useless.
Not that the trials and tribulations of being a journalist should concern anyone working in tech, but since Big Blue is so desperate not to talk about Db2, I thought it would be fun to give the notoriously reliable database another airing.
Db2 was modeled on the ideas of IBM researcher Edgar Frank “Ted” Codd, who first described the theory of relational databases in 1970. The first products became available on IBM mainframes in 1983 and later on Unix, Linux, and Windows. The product has been styled as DB/2, then DB2, before settling on the current Db2 in 2017. According to database pioneer Michael Stonebraker, it was Db2’s adoption of SQL on its world-dominating mainframe platform that led to the query language becoming the de facto standard.
Since the 1980s, Db2 has become a database of choice for applications that need to be both big and dependable. Even today, among Db2’s user base, banks make up nearly 43 percent, according to some measures. Among them are American Express, Bank of America, Citibank, and Deutsche Bank.
Experts say Db2 is well engineered, reliable, and not as dated as one might imagine. Its only commercial rival, Oracle, has hardly updated its underlying architecture, but that doesn’t stop Larry Ellison’s visage appearing from over the horizon every time it gets a makeover.
So why does IBM not talk about Db2? To be fair, it does squeeze out some news if you know where to look.
Last week at a North Carolina International Db2 User Group (IDUG) conference, IBM promised Db2 LUW v12 would include an AI-powered query optimizer, which would “allow Db2 to continuously learn from customer’s queries and achieve up to three times query performance improvement over prior version.”
In November last year, Db2 became available as an AWS Relational Database Service (RDS), joining a family of databases on RDS that also includes PostgreSQL and MySQL. The announcement, however, came from AWS, not IBM.
The move followed a statement that IBM would attempt to execute a “cloud-first” strategy with Db2, although that too came from an IDUG conference.
However, given the context of Db2 users supporting applications vital to their business, maybe IBM could have told the world more about their future in the cloud. The cloud, after all, is inescapable as a tech trend. Even if IBM talks directly to Db2 users, any signals from IBM about it outside of this group might give the wider tech world some confidence in its future.
IBM may have its own reasons, but since it is seemingly unwilling to share them, we are left to speculate.
The tech industry is led by investors, rather than customers. And they expect growth – not just normal, steady, above-inflation growth. They expect hypergrowth, as characterized by former Snowflake CEO Frank Slootman in his book Amp It Up. The tome is all about hypergrowth for companies that focus on a “land and expand” strategy, targeting a niche market then bursting out into adjacent markets.
It sounds great on paper, and early Snowflake investors were undoubtedly happy with the concept when Snowflake achieved a $120 billion valuation not long after its 2020 IPO, when it was briefly valued at more than IBM.
But Snowflake’s flurry did not last. Its shares slid from the end of 2021, and it has barely made any gains since. Not that that would concern Slootman, who left the company in February, leaving investors wondering what sales forecast are based on now that its biggest customers have cut spending by optimizing deployments.
Db2 is so well established in its market that it could never exhibit Snowflake’s kind of growth, yet it still sits just two places behind Snowflake in the DB-Engines ranking (ninth place).
Compare Db2’s fate with another relational database with long-established roots. Postgres was launched not long after Db2 in 1986. After a promising commercial edition stumbled, open source PostgreSQL steadily grew, and today it enjoys a commanding position as the leading database for professional developers and fourth place in the DB-Engine’s rankings.
Not only is PostgreSQL successful in itself, it has also become the front end for other database services for the larger cloud vendors, and companies such as CockroachDB and Yugabyte. Some commentators see it becoming a de facto standard front end, while vendors innovate on the back end. That ship may have sailed for Db2, and IBM might rue the decision to let others tell its story. ®
