Ask TPM: Two System i Questions, and Two Responses
October 29, 2007 Timothy Prickett Morgan
Our experts over at Four Hundred Guru get asked a lot of questions and they provide a lot of good answers. But some of your questions are of a strategic or tactical nature, having to do with broader issues in the information technology space or with the i5/OS and OS/400 platform. So if you have a question you want to get my take on–or you want to use me to get a straight answer out of the people in IBM who know what is going on–then send me an email and I will see what I can do. Here’s a recent set of questions I received from a reader of The Four Hundred.
Good morning. I really enjoy reading your in-depth columns. Here are my questions.
First, I have relied on the AS/400–ok ISERIES, er System i–since 1990 to make a living. What do you see happening to the platform, say over the next five to 10 years? Is it possible to break it down farther to equate this to jobs as well, like system admins and programmer/analysts?
Second question. What type of voice do software vendors have with the decision-making processes that IBM has?
Multipart questions are fine with me, Jack.
Like you, I have been relying on the OS/400 and now the i5/OS platform to make a living since the debut issue of The Four Hundred in April 1990, a newsletter I started working on back in the summer of 1989. I didn’t know as much about computers or writing as I thought I did back then, and I admittedly did not know the first thing about journalism, since it was not a career path I really contemplated. It is a job I fell into, thanks to necessity and recession. Come to think of it, that’s also how IT Jungle got its start in 2001 as well.
With IBM reporting declines in quarterly System i server sales for the past eight quarters–and double-digit declines in seven of those eight quarters–it is natural enough to be a bit freaked out. But IBM’s ability to sell new iron is not necessarily an indicator of the health of the overall OS/400 and i5/OS ecosystem. To be sure, If IBM’s sales were growing and we had anecdotal evidence that new customers were coming to the fold, consolidating lots of other platforms onto the box and adding new applications, too, we would all be ecstatic. (This is essentially what has happened to the mainframe base in the past two years, until it hit a wall in the third quarter of 2007.)
The last time I did the math, with a whole lot of witchcraft built into it comparing the server ecosystems of the world (in 2005 using 2003 and 2004 statistics, see the Related Stories section below for more on that), I figured that the OS/400 ecosystem–including annual sales of servers and storage, systems and application software, development tools, maintenance, and other services plus the value-adds that third parties reselling it take–accounted for $6.5 billion a year; this ecosystem was declining slightly on an annual basis, perhaps at a few percent, while the overall server ecosystem, at about $365.5 billion, was growing at a rate of a few percent per year. Two decades ago, when we both started out in this base, all of the iron and systems software was ridiculously more expensive, so even though there was a much smaller base, the OS/400 ecosystem was larger. In 1993, I figured the OS/400 ecosystem was around $12.5 billion. It would be easy to jump to the conclusion and say that by 2012 or so, the ecosystem will drop again to $3.5 billion. It might, but a lot of the ecosystem comes from suppliers who sell gear to the 200,000-plus companies worldwide that have the OS/400 and i5/OS platform installed and most of whom have no desire to move off the box for a wide variety of reasons. We have already lost around 80,000 customers since the AS/400 base peaked in 1998, and I think that the rate of exit off the platform has slowed considerably as the economy has improved.
Against that backdrop, let’s talk about the future of programmer/analysts and system administrators for the platform. If you want to be generous with the numbers, there are probably 215,000 OS/400 and i5/OS companies in the world, and they tend to have just one coder, or maybe two or three coders, do double time as analysts or administrators. That’s about 645,000 programmers in 2005 against a worldwide population of 11.6 million programmers (based on information from Evans Data), or about 5.5 percent of the programmers in the pool who do RPG at least some of the time. If this RPG pool remains about the same and the worldwide programming pool grows to 17 million by 2007, RPG drops to 3.8 percent of the base. Just to stay in the same place, RPG has to grow to 940,000 programmers. I don’t think this is going to happen. What I do think will happen is that the i5/OS and OS/400 platform will increasingly become the nexus of many different programming languages. Gartner analysts expect that by 2008, over 80 percent of new applications will be coded in so-called “managed runtime environments,” such as C#, PHP, and Java, up from 60 percent in 2006. So long as power and cooling issues in the data center do not require us to start using compiled rather than interpreted languages, the move towards these languages seems inevitable. (It also suggests that IBM should create a portable RPG runtime environment that is abstracted from the DB2/400 database and the System i iron.)
With Web services and services oriented architectures becoming more popular, it is possible, if the economic case could be made to have the System i to be a hub in an SOA-style hybrid application that spans platforms. That might mean companies need fewer hard-core RPG and COBOL programmers and more programmers who know more about SOA tools. IBM certainly wants i5/OS shops to embrace its Enterprise Generation Language, which it is positioning at the heart of SOA applications and which it wants RPG shops to develop even though the Rational tool behind EGL only kicks out COBOL and Java executable code.
Given the tendency to use virtualization on the server and to have more automated server provisioning and administrators managing more servers, it seems likely that there will be fewer system administrators at large i5/OS sites five to ten years from now. But among the smaller shops, which dominate the i5/OS and OS/400 installed base, this automation will not mean a loss of jobs but relief from certain responsibilities. Companies will still need a programmer, after all.
The real question is what kind of programmer will they need? If I were an RPG programmer and I wanted to make myself useful, I would be studying PHP and MySQL, since there are so many applications based on these two technologies. If core business applications will be coded in RPG, the way that they interface with the Web will be increasingly dominated by PHP (no one really wants to learn Java), and for separate Web applications, MySQL is often the database of choice. The other thing that I would do is get some familiarity with the legacy application modernization tools available from a slew of vendors, which move code out to the Web and into the SOA world without the kind of hand-coding that a hybrid RPG-PHP approach will require. Some companies will be too cheap to pay for these legacy modernization tools, and others will only code with them and will dissuade their programmers from trying to reinvent the wheel.
To put a number on it, I think that there will be slightly fewer RPG programmers in the world–maybe 500,000–a decade from now, and that RPG programmers will be using other tools for more of their day than they currently do. Across the base, there will be a mix of Java, PHP, and a plethora of 3GL-style tools that speak (heaven help me for using this idiotic acronym) Web 2.0. There will be fewer pure system administrators across all server platforms thanks to automation, and among i5/OS shops, the workload will get lighter as system administration tasks are done by the system itself. Luckily, as the current crop of RPG coders and i5/OS administrators starts to retire, there will probably be more job opportunities than people to fill them, since skills will be scarce and demand will be steady. No one is training newbie RPG programmers, no one is training System i administrators, and demand is not going to decrease by all that much–provided that there is not a big recession that leads to a depression.
If that happens, heaven only knows what happens to any customer base. Recessions drive technology shifts, and the next one could be to hosted applications where no one has a programmer, but rather they have a shared application backed by shared programmers.
As for the input that key independent software vendors have on IBM, all I can say is that they have about as much input as we appear to–which means very little–and they suffer from IBM’s actions and inactions–just like we all do. I think IBM spends a lot more time listening to what key resellers tell them about exactly how high it can push prices and how little performance it can put into a box and get away with it. This wrings the most profit from a system sale, and if the ISVs benefit from that, then IBM is happy to let them win, too. But the system is no longer designed to assist ISVs who peddle applications, and you need only look to the fact that there needs to be a specially priced Solution Edition tied to a very specific Oracle and SAP suite. While I applaud the fact that IBM is willing to give discounts to Oracle and SAP customers to win deals, it used to not be necessary to split the product line up. The Application System/400 was for all ISVs, for all OS/400 shops, and it embraced RPG and COBOL joyfully and proudly. And it was a very competitive box in its time, and hence it came to dominate the midrange.