Successful, profitable businesses run largely on legacy production systems. Many of these systems are mainframe-based. Take the legacy away and any business based on transactions will suddenly disappear.
As someone that hopefully understands the reality of enterprise computing, having spent 17 years in the business, I find a lot of industry bullshit about innovation vs maintenance spend pretty wearing. Hundreds of vendors have pitched me to explain that we’re all doing it wrong. Gartner Group says 90% of IT spend is maintenance which is a bad thing. McKinsey argues organisations should only spend 40-60% of their budget on maintenance. The figures then get repeated ad infinitum, with the subtext being the maintenance is bad and new app spend is good. Its kind of weird, but mirrors the wider culture’s cult of youth, where female newsreaders are not allowed to have wrinkles, and our most feted CEOs are seemingly all under 30. Yet the part of the business making the money is dismissed.
It was extremely welcome therefore to see the CEO of RBS admit that the recent massive outage at the the bank, and its subsidiaries NatWest and Ulster Bank, was because of a failure to invest sufficiently in legacy.
“RBS has seen a big mushrooming in spending on technology. With hindsight maybe a bit more of that increase in spend should have been in the core, taken-for-granted systems that work every day.
Turns out that new systems of engagement are important, but you neglect existing systems of record at your peril. Talking of peril, the issue of IT-related bank outages has now come to the attention of the Financial Services Authority, with the Financial Times reporting that maintenance is now a regulatory issue.
Now let me be clear here. I am not saying legacy is always good. On the contrary legacy is only valuable if you continue to invest in it. You need to pay down your down technical debt. Investments in legacy are not investments in nothing, they are investing in the walls, floors and roof of your business.
If you live in a house and never invest in maintenance, or have a car and never go to the garage, your investment will fall apart expensively and potentially catastrophically.
The FT story goes onto say that “IT specialists say many smaller lenders and those operating in emerging markets are leapfrogging established western rivals in technology performance as they build new, unified systems.”
Aha – the bullshit arrives. Who are these “IT specialists”, or as Computerworld calls them “analysts”? What- they were afraid to be named? Lazy journalism I fear.
The truth is Asia has been a strong growth market for IBM’s mainframe business, precisely because of the scale challenges of providing transactional services to countries with massive populations – I call it Scaling the Great Wall of China. Asia, which is rather less youth obsessed, wants to run its banks on the same systems we use, rather than betting on something untried.
Mainframe shops that invest in getting current with operating systems and software can run mainframes at lower cost than those that don’t. Investing in legacy is, and can be, a competitive advantage. Is legacy always better than new application build-out on new hardware systems? Certainly not. But neither is migrating to the newest cool shiny thing a guarantee of lower cost of management and operations. Mainframe shops running current hardware and software can take advantage of lower cost subsystems for running Linux, data and Java workloads for example.
Bottom line- not investing in legacy is a false economy.
Teressa Jimenez says:
August 1, 2012 at 1:55 pm
James – very nice insight. It is the marriage of continued investment in legacy – maintaining it and evolving it which is the point. Stability and wealth are built on these principals. Thank you for calling this out.
jgovernor says:
August 2, 2012 at 5:29 pm
my pleasure teressa. and thanks muchly for taking the time to comment here. not enough IBMers do! 🙂
Martin Packer says:
August 3, 2012 at 9:44 am
Of course I like this. 🙂 And it echoes much of my own thinking. I see a lot of this lack of attention to maintenance – or even understanding what they’ve got.
Ray Mullins says:
August 3, 2012 at 4:28 pm
Amen, Martin, amen. If you buy the snake-oil of outsourcing overseas with the crack cocaine of cutting costs, your IT becomes out-of-sight, out-of-mind. Then your mission-critical processes come to a collision with a brick wall, and suddenly you remember what you outsourced.
There’s a reason why it’s so cheap folks – the people they are hiring don’t even come close experience-wise to the ones you are letting go, you know, the ones that keep things running and making it better.
DancingDinosaur says:
August 3, 2012 at 12:23 pm
Excellent piece, James. Those who urge innovation over supporting the existing infrastructure fail to appreciate how easily IBM has made it to innovate on top of the mainframe through Java, WebSphere, Linux, IBM connections, and a host of other tools and technologies that work with the mainframe.
Mainframed767 says:
August 3, 2012 at 1:55 pm
Why do we even call mainframes “Legacy”? Linux has been around since the 90s, same with Windows. The server architectures get incremental upgrades, sure, but thats the same with the Mainframe. Hell, IBM released the newest version of z/OS in september. Calling it ‘legacy’ shows a complete misunderstanding by tech journalists of how dependent the global IT marketplace is on mainframes.
How would this article have gone if instead of the mainframe it was Windows NT running on Pentium IIIs?
jgovernor says:
August 20, 2012 at 4:39 pm
angry much mainframe767? when i was a tech reporter my nickname was “legacy boy” because i covered mainframe and more mature tech stuff. legacy really just means a production system, but it also has a common and widely held definition in the market. and yes windows NT is indeed legacy now- a smart organisation is going to invest in their windows server portfolio too.
M-A-O-L » Invest in Legacy, Invest in the Mainframe says:
August 3, 2012 at 5:23 pm
[…] via The Lesson of RBS: Invest in legacy, Invest in the mainframe. […]
Vince Smith says:
August 10, 2012 at 4:21 pm
All that glitters is not (necessarily) gold! Good point well made.
MikeC711 says:
June 25, 2013 at 2:30 pm
I’ve been on software companies that had a “mainframe” runtime that generated 80% of company revenue and the sexy systems of engagement where they figured “the future” was. The staffing was the reverse of the revenue (80% sexy distributed stuff, 20% “mainframe”). The company eventually folded. Even with this example though … Systems of Engagement should not be belittled and Systems of Record should be stable and monitorable enough that maintaining them is not horribly expensive. As new work comes along … there are metrics to identify where the various platforms excel.
I applaud RBC for seeing the problem and not doing the knee-jerk reaction of “migrate off the mainframe to the “modern” and “cheap” platforms. Sometimes migration is a good idea, often it is not, and never are the migration consultants up front about all real costs. As for “modern” … Z can run the latest versions of new server software, and when I say “cheap” … it’s the old story where people don’t factor in that that 32 core “cheap” box needs all the software licensed on 32 cores (and the 32 cores for redundancy and the 32 cores for DR and the 16 cores for dev and test, …). There still truly are times when migration makes sense, but a large % of migration projects would love to go back and in time if they could, as that 2 year ROI turns into a 29 year ROI (or worse).
Derek Britton (Micro Focus) says:
July 4, 2013 at 10:42 am
I would agree with many comments here as well as the article itself. Another gem, Mr. Governor.
The notion that legacy is fundamentally a bad thing is too simplistic a view -a well-cared-for core system will continue to provide business value that far outweighs the cost and risk of replacing it. Moreover, age is not a sensible indicator of value. One only has to think of recent technology “standards” that have gone as quickly as they appeared to witness the folly that everything new is the right option. As ever, a balanced view needs to be taken.
Investing sensibly in improving the operational efficiency of IT (and therefore the business) is the right driver, which will mean making investments in existing systems as well as any newer innovations. In fact, seeking to deliver and manage existing workload more efficiently negates many concerns about so-called legacy by removing needless costs, improving core system quality and mitigating the risk of failure. Using emerging standards like Eclipse, zEnterprise and other 21st century innovations allow tried-and-trusted mainframe technology to serve the new demands of today’s market more efficiently.
jgovernor says:
July 4, 2013 at 2:56 pm
thanks Derek. I think Vince in particular nails it. We apparently celebrate go faster stripes, and fancy fibreglass spoilers, rather than a V8 engine.
Derek Britton (Micro Focus) says:
July 5, 2013 at 7:42 am
So true, don’t we all. It comes up a lot! http://tinyurl.com/q2497yv
Matthew Malin says:
August 1, 2013 at 2:22 pm
Thank you for the excellent article. Some recent work that I’ve done have included modifying and squeezing the most out of 30 – 40 year old mainframe applications that are still valuable rules engines. They are paid for and stable. So maintenance is minimal. I am currently in a shop that is switching over from stable mainframe technologies to open-source consortium/collective created Java and Oracle based applications. The IT outsourcing has never paid for itself. With high rates of turnover business knowledge is never obtained/retrained. There seems to be no understanding that knowing COBOL/CICS/DB2 is a quarter or less of what a developer/analyst needs to know. Again thank you for an excellent article.