Readers React to Project Prometheus System i Effort
June 26, 2006 Timothy Prickett Morgan
|
As you might expect, there has been a lot of talk about the Project Prometheus effort to rejuvenate the System i platform since I did my initial story on the secret project two weeks ago. Many of the things that people said in their comments to me about the project were the same kinds of things we have discussed over the years about how to differentiate and peddle this platform. I find the consistency in our views comforting, even if I find the consistency of IBM‘s short-sighted, self-centered behavior–how shall I put this?–freaking annoying. I am a firm believer in serendipity, and I sometimes think that synchronicity is real. The former is a fun way to learn about the world, and the latter is just plain spooky, but often seems to be a true force in the universe. It is perhaps not a coincidence, then, that a few days ago, I was reading the May 8 issue of The New Yorker magazine, in particular a story entitled Circle of Fire. The author, Andrew Solomon, wrote what I assume is a brilliant piece on Libya, with all of its contradictions, problems, and hopes as well as “the Leader,” Muammar Qaddafi. The fire that the article refers to is Qaddafi himself, and the idea is that if you are not part of the circle, you are out in the cold; and if you are close enough, you can warm yourself. But if you get too close, you get burned. A few sentences made me laugh out loud, and they are relevant as we ponder Project Prometheus. “Officials in Libya seldom say no and seldom say yes,” Solomon explained. “Libyans use a popular Arab term: IBM, which stands for Inshallah, bokra, moumken, or ‘With the will of God, tomorrow, maybe.'” You can’t make the good stuff up. It just happens. I have offered a lot of advice to IBM over the years in the pages of this newsletter. Some of it seemed to stick, but most of it doesn’t. But, my job, as I have always perceived it, is to persevere, to think of options, to present criticisms, but also hope. There are always possibilities, a wise man once said. In this sense, I have felt not like the legendary Prometheus, but more like the existential Sisyphus, condemned to push his rock up the hill to only see it roll back down again. I think we all feel a bit like that, just being alive. But I have always believed that when Sisyphus wasn’t grunting from his labor, he was smiling, and when the gods weren’t looking, he actually sang. What a gift it is to know every day what you need to do, and to put your back into it. I will let you give IBM your advice on what to do with the System i in this story. I have some advice for the people behind Project Prometheus: Never take credit. Never say you are officially doing it. You’re like a superhero with a secret identity. Things just show up when people ask for them. Work through proxies. Build mystery and maintain it. That’s part of the story that will evolve. Think of yourself as The Lone Midrangers. Expand out and embrace the entire community. Get them to help and tell them to keep their mouths shut about where the resources come from. Just do what you know needs to be done. Anyway, without further ado, here is a sampling of what some of you thought about Project Prometheus. Some of these had to be edited–this is a family newsletter–and some I could not even print at all. Some comments–and I have no idea how many–only went to the alias mailbox we have set up at IT Jungle, which is on our Contact Us page under Prometheus, Rebel with a Cause. “System i” HA! I have been dedicated to this platform for the last 15 years. I have founded two great software companies focused on the AS/400. As an ISV, I have made IBM a ton of money, and never received one lead for our horizontal software product. I do not have the energy any longer to discuss/explain/excuse/heckle the NAME of the box with our prospects and customers. We have made a corporate decision to call it the Application System 400–AS/400. Period. I noticed Timothy using the term OS/400 very specifically… The renaming has done more to hurt this excellent computer than I ever imagined as I gasped at the introduction of the iMbecilic iSeries name. Now it only gets worse. It is a damned AS/400. Forget running Windows and Linux and consolidating the crappy little servers that eat your lunch, it is an Application Server, and one of the BEST. We are RPG programmers, and PROUD of it. Yes, we do Java and C and PHP and all kinds of stuff, but our heart is in the native AS/400. I can live without leads from IBM. But I cannot handle their constant efforts to negate the platform. And I am now starting to tell everyone who will listen. It is an AS/400! –Name Withheld First, I have to say that the shop where I work lost the iSeries battle and we are now cutting over (about half the way through) to the HP-UX server running Oracle Applications on an Oracle database. And I must say every time I turned around, we lost functionality we had running JDE World on iSeries. There are two reasons why we made the switch away from iSeries:
1) The CIO wanted to run on commodity hardware and a common database (read Oracle) No matter what we or IBM did, we could not convince her of the value and magic of the system. In fact, at one round table meeting she said this sounded like an OS religion discussion and she would never run AIX in this shop (even though we did at the time). What could be done to battle this mindset with the people in power? [Decision makers, not the power processors 😉 ] Well I think there are a number of things that could be done. 1) Run Windows, run Windows, run Windows. Let’s face it, Windows has won the datacenter and to stay relevant we need to play by running Windows in an LPAR. I don’t care if it is slow or in emulation mode or what, but the i5 has to be a “run everything” box and running Windows is part of the picture. 2) Natively support Windows applications. Not only run Windows, but also run the Windows applications without the OS. 3) Natively support ASCII in DB2/400. DB2 on i5 has to perform just as well in an ASCII mode as it does in EBCDIC. 4) Demonstrate that Power processors are faster at running applications on Linux and Windows than Intel and AMD processors can. Better yet, faster even in emulation mode. I have searched the Web and I cannot find any data that shows that Power processors are any better than Intel running Linux or OS X or anything else. (I have found that Windows runs faster on the same Intel hardware from Apple than OS X does.) I think for the i5 to stay part of the homogeneous data center environments, it must run Intel/AMD operating systems natively, emulate Intel and AMD processors, and do this in an LPAR. Otherwise, it will continue to be the hardware that does not run Windows and Windows applications as well as other niche Unix apps (like an Oracle database). Until the i5 can run Windows and an Oracle database, it will hold on to that legacy label. Because the big ERP applications no longer run only on the iSeries, but also run on other platforms, while the i5 can only support some of the other platforms work loads. The i5 is not commodity hardware, but it can’t natively run the stuff that runs on commodity hardware and therefore does not warrant a second look. Best wishes, –Michael I work for a Fortune 500 company. (I won’t tell you the name because that might get me in trouble.) It started on the System/38 in the late 1980s. I don’t know exactly when, since I came to the company in the mid 1990s. Multiple AS/400s carried the bulk of the corporate, back-end processing, and still do. But, their days are numbered and we are now on a multiyear project to retire the platform. How did we get to this point? There was a desire to get into GUI front ends and this invited in PowerBuilder and VisualBasic. (Of course, this being an unstructured environment, we now can’t find some of the source code.) IBM failed to give us a “real” programming language until RPG IV Free came along and most Unix and Windows developers really struggle with columnar languages. I know–I work with a lot of them. SQL’s poor performance in the early days, the additional cost of the SQL development package, and a lack of understanding of the database caused us to write some poorly performing, native I/O-based, inflexible systems. It took years to get journaling and commitment control accepted. A lack of strong leadership, and a series of upper and middle managers who didn’t understand what the machines could do, opened the doors for Sybase, SQL Server, Windows NT, Unix, DB2, Oracle and a few other software/hardware combinations few people still remember. Top management changed and realized we had way too many languages supporting way to much duplicate data in way too many silos on way too many machines and server farms. Consolidation and “legacy” replacement became the new, and I believe correct, direction. IBM is managing and staffing a couple of major projects for us. In all its proposals, the iSeries was never mentioned. In fact, I have met only one of the myriad IBM employees who knows anything about the platform. Most of the big components of the new systems will be running on pSeries machines with AIX, with Wintel boxes and Linux in smaller sites. IBM will be hosting and managing the pSeries machines. I believe that from IBM’s global perspective, Unix, Linux, and Windows boxed are much more lucrative in terms of hosting and support, and such service is where IBM sees profits these days. Therefore, IBM is not likely to push the iSeries in any proposal because it can’t make enough money on it. In our company, the iSeries is slowly dying. I really don’t see IBM making a huge effort to keep it alive anywhere else, because it isn’t profitable enough to them. It won’t die slowly, and it won’t die quietly, but die it will. Regrettably, I think the battle to keep alive the best business machine that IBM ever created is already lost because it is not in IBM’s best interest. –Name Not Given If I understand what your group is undertaking, I have had the same thought for years. If IBM is not going to market the platform, turn it over the people who will. If I could feed my family being an iSeries evangelist, I would do it in a heartbeat. I am not salesman, mostly because I cannot sell something I don’t believe in, but the iSeries sells itself if we could get the truth out about it and the hype of the other platforms could be neutralized. I said that to lead to this we need a way to cost justify the iSeries against the other platforms. What would be a great tool is to take the CPW of an iSeries and show the equivalent number of servers for Microsoft or size of the machine for Unix or big iron (do I dare say. . . mainframe) offerings. Also adding in factors of variable cost: How many VTAM admins for an equivalent machine? How may DBAs and system admins for Unix or Windows platforms? This needs to be backed with real life examples so the proof is not statistical pie charts, but living examples that the bean counters can’t dismiss it as “marketing hype!” I am in a shop right now that has just that kind of living proof , but IBM would not touch it with a 10-foot pole. We came off an IBM mainframe and have saved hundreds of thousand of dollars. The issue now is fending off the Windows camp. Without hard proof, that camp’s logic is “see, we can save thousands more” and once you start down the dark path of migration to the next platform it very hard–nigh, next to impossible–to return even if you prove them wrong! The issue is IBM has its loyalty too split to take this kind of a stand to help out with this plight. It’s like IBM’s saying “The other brands might be offended . . . It might seem like we are prefer our own brand instead of another’s . . . that older technology might lose out if we put it to real competition.” How would this fly as a mission statement! Sorry about the rant, but someone has to take a stand. I hope that is what your group is planning on doing. Else the iSeries may be another interesting footnote of a great idea that just didn’t quite make it and IBM may be subservient company feeding of the crumbs someone else is willing to throw them. If there is anyway I can help let me know! Let me know if you can help me! –Bill Dear Prometheus Team I don’t know who you are, but I bet many of you know me. You know the investment in time and effort I have made and my firm has made to live up to our part in being part of the community of AS/400, iSeries, System i. First, the name means nothing, stop messing with it (its past is the System/36 and the System/38, so this isn’t too far from that) and stop making it the biggest issue you have to address. Imagine the time and $$$ spent trying to change it, market it, and defend it. Stop the bleeding. You’re right in that TV ads don’t sell systems, you don’t see too many System z ads. Windows systems rely on ads to build upon brand name recognition–but they are software not hardware stop selling against it and provide the best possible platform to run your Windows servers! Yes, you need to support the trade publications that your existing and potential new users are reading so the name isn’t new, but CEOs don’t read those magazines. What are you doing to reach out and touch them? What does the System z team do to market, where do they market, how do they connect with new clients? I just learned during a road show that 70 percent of IBM’s customers have a System i or predecessor system. Why don’t you tell the world that so everyone stops wondering if this machine is coming to end of life? You have 100 percent market share on MARS. Promote that. (A little humor doesn’t kill you.) In the 14 countries I have been fortunate to travel with IBM and promote the System i, only two countries had local technical sales support (Canada and USA); all others relied on remote support back to the aforementioned two countries. When we bought our first AS/400 B35 back in 1989, we bought it not because of the machine, but because of the software that ran on it. You need to make sure you have software to cover the broad-based needs out there. Sell on solutions to new clients not hardware, and reinforce with existing clients as you have been doing the reliability, flexibility and great consolidation/infrastructure simplification platform that it is. These are a few of thoughts of the top of my head, you know I am always available for you, don’t be afraid to call!!! Sincerely, a loyal iSeries Customer –Name Withheld Here’s what I need: User-based pricing. Enable me to write and sell an application to local doctors, lawyers, accountants, libraries. An application that would be used by one to 10 people. i5/OS has to be priced per user. A yearly subscription is best: $200 per user per year. Merge the i5 and the p5. i5/OS is then able to run in a partition of the millions of p5 systems in use today. How is that for marketplace awareness? Improve the programming languages and runtime of the system. RPG is primitive compared to VB.NET and C#. .NET is much better than ILE. This is very bad. IBM the company falls short when it comes to programming languages. Where is the IBM equivalent to Microsoft’s MSDN magazine? Stop gearing down your hardware. Modern languages and runtimes need a lot of CPU cycles. If i5/OS is sold per user, IBM can open the CPU floodgates without concern that 200 OS400 users will be able to run on a $5,000 rack-mounted p5. Work hard. Please save our system! –Steve RELATED STORIES Happy 18th Birthday, AS/400; Time to Leave the Nest Project Prometheus: Playing with Fire |
