Slashdot Mirror
Selling Linux to AS/400 Shops?
cgh4be asks: "I work for an IBM business partner and each year we host a technology seminar for our customers, most of which have AS/400 backgrounds. I am the 'Linux guy' at work and have been chosen to give a 1 hour presentation to these IBMers about linux. So, my question to the Slashdot community is, what points and information should I touch on in this presentation to make it effective? I'd like to give some history about it of course, but the goal is to get them to start using it in their businesses."
My question is: why?
AS/400 are decent machines with a good operating system. It has the virtue to "just work". I'm not really sure you would want to replace an AS/400 with a Linux box. The AS/400 is -after all- a proven and stable platform.
I'm all for Linux, but not in this category of hardware.
You must believe that Linux is a better choice for some reason. You're going to tell them that this young, haphazardly developed, upstart OS is a better choice than a mature, thoroughly architected, industry standard OS.
I think you'd better tell us, rather than the other way around.
I have been pwned because my
AS/400 units are database servers. Sure, you can get an web server, and a firewall on the unit, but they all run out of a database anyway. The neat feature of AS/400 units that I like is I can ftp to a directory, and upload files, and they are added to the database.
Linux is much better in the other services that AS/400 units don't cover very well, but that they can cover. Such as firewalls, mail servers, file servers, and they could be used as a way to interoperate with winblows... oh, don't forget... if you get db2 for linux, you can run your database on linux and do an offsite remote sync up.
Our AS400 systems were very impressive-looking in a "I AM BORG" type way. They had great reliability and service, but they were older models, expensive to maintain, and didn't have very good performance. I suspect other shops are in the same boat.
:)
We started using Linux machines as development platforms. The AS400 guys could run db2 and other IBM technologies on Linux, which was a big win. The development stations were dirt cheap and for many jobs faster than the AS400.
Eventually as that department became more comfortable with Linux they stopped placing every business application on the AS400s if a properly configured Intel/Linux box could do a cheaper/faster job without unacceptable risk exposure. It turned out quite a few applications fell into that category.
The key is to have someone who has a deep understanding of linux and also can talk to the hardcore "mainframe" admins. There is definitely a culture shock to overcome. Imagine trying to explain Linux to someone who calls service for any task that might require opening a box, pays 10k/year license fees for a glorified tar backup system, and knows so many esoteric commands and proprietary-os arcana to make VAX and UNIX together look like kiddie toys.
Our AS/400s are actually new. The oldest one is only a year old. For us, despite our deep Unix and NT backgroud, Linux on an iSeries (there, I said it... the new name... blech) holds little more than sideshow value. For us, we have almost no investment in AS/400 admins or developers. We just add HP/Compaq servers if we want more Linux.
/400 shop. Someone like our parent company that has AS/400s everywhere (30+) probably sees a huge benefit to running several Linux virtual servers on an LPAR* or older single system. The key to stress is reuse of older systems. Some of the first generation RISC systems (F10s and the like) would make great highly-available Linux machines.
However...
We aren't a typical
What's more, a company with a large investment in the hardware has probably moved all but some development off of a depreciated system. The shops that have two systems still running V3R7 won't be interested.
Part of the problem is that IBM doesn't really push economical LPAR-ing for Linux up front. In our case, they said "Sell you on LPAR? Well, if you aren't interested, we can't change your mind." Really, the Rochester salespeople are technically very knowedgeable, but they just don't "sell" sometimes. And, I can't go back and get the money now to do an LPAR "just for Linux".
The other issue is that IBM (to date, anyway) doesn't push AS/400 to SAN connectivity, which increases the value of the AS/400 as the key system in your enterprise whether you run Linux on an LPAR or not. You could, in theory, run a fairly-sized enterprise with a handful of 8xx-series machines. You'd have an NT-based FSIOP (or Novell, if you swing that way) if you had a need for it. You'd LPAR a big box to carve out several virtual Linux machines to keep your development options open. And, you'd have great MTBF on the hardware (except for the problems with the 17GB drives... those die like flies). Back all of that with an enterpise-class SAN (Shark, EMC... no Hitachi or HP XP-series because they don't support the 520-byte block size), and you can snap whole systems from LPAR to LPAR and system to system. Then, you can use a CNT-like solution for data replication over IP to a disater recovery site because all your data is on your SAN.
Sorry for the ramble, but I guess the point is that making the case for Linux on an AS/400 is part of a big picture. I'd say the cost-savings of having a handful of the same hardware in your computer room but the flexibility to run virtually any application is an outstanding point.
* LPAR == Logical PARtition... it's similar to how a Sun E10k can be carved into separate systems at the hardware level. And, yes, lots of other companies have done the same thing. It's not new.
Amateurs discuss tactics. Professionals discuss logistics.
Specialization, specialization, specialization.
Oh. And reliability.
So four words.
I cannot begin to describe how bullet-proof an AS/400 is. People simply don't believe me. One poster said that he knows people who IPL once a year. That's probably because it's *intended,* by the way - not because the machine hiccups. These suckers are the equivalent of cockroaches - they'll survive a nuclear blast.
But other than reliability, you're going to have to go after custom software markets. And not the glamorous stuff - the boring flat database stuff. I work for a good-sized printing company. We used an industry-specific ERP on an AS/400 for years and years. We still use that same app - only we now use the Windows version.
The AS/400 version was far less maintenance than the Windows version for various reasons - but the bottom line is the specialization. This isn't a couple of spreadsheets and a bar code wand out on the shop floor. This is much more involved than that and my industry type isn't the only place where this kind of software shows up. At the AS/400 mini-conferences I attended there were casino IT people, tool and die shops - even the car dealer I frequent has an AS/400 running industry-specific parts-tracking and job estimating software.
Plus these are not necessarily companies that have large R and D or programming departments. Nor *should* they. There are a lot of companies that are similar in size to ours that have IT people, a geek or two, but no one that really has the time, wherewithal or desire to replicate industry specfic software that can be easily purchased (for around $20,000 for all of the modules in our app) and supported (for about $2,000 a year.) Yeah, the hardware is expensive, but so is my time or the time of a programmer to create something that is customized for us. And oddly enough, I do have other things to do...
If you were selling Linux to me as a replacement for an AS/400, I'd first ask you what software is out there that will run my kind of manufacturing plant. I've already been down the "take some general software and customize it" road. It's a dead end.
If the answer is that I have to hire someone to write it or come up with something on my own or that you'll gladly look into consulting with my company on our specific needs, I'll hang on to my Black Beast, thankyouverymuch.
Consigned to flames of woe.
WRKSYSCTX is "work within the system context"?
WRK = work
SYS = system
CTX = context
All of the OS/400 commands are like this. Three character 'words' combined in strings of one, two, or three 'words' as in your example.
There were exceptions. WRKSPLF for example,
WRK = work
SPL = spool
F = file
Note that WRK is always work. This is true for all of the three character codes. This made the commands easy to learn and remember. Contrast with *nix, where each command name has to be learned separately.
And it is not in condensed combinations of verbs, it is verb noun modifier.
As to the all caps requirement, are you sure that wasn't just a convention in the document you were reading to distinguish commands from the other text? OS/400 has no such restriction.
Steve M
In addition to many of the good recommendations you are getting below I'd recommend:
Unix for the Mainframer
While this is more directed at MVS and not OS/400 the first chapter could end essentially being your 1 hour talk. A great example is the difference in Unix speak between: command, script and application. The underlying philosophy here is one of an interactive environment where small non-job specific programs (commands) are put together for job specific tasks (scripts) and that these are distinguished from customizing large systems (applications). Just explaining the pros and cons of this model will be insightful and will take up a good deal of time.
Incidentally in terms of the cons as a Unix guy you may not have thought about these. Here is a pretty good anti-Unix pro mainframe article which will help you see the other side S390 vs. Unix servers.
Anyway hope this helps