System iNetwork

Maybe I am pessimistic in my thinking, but let me play devils advocate for a moment. After iManifest garners all this attention it is trying to win over, what then do you tell people that become interested? That they can run PHP and Java on a platform that is stated to be better than Wintel/Lintel?

I would have to wager and guess that most are won over with the IBM i through time (at least I was) - after-all, how do you convince somebody that library lists, job control, job logs, call stacks, and integrated databases are a good thing? Or convince them that proprietary actually works in their favor (for those that are bent strictly for opensource)? Or try to convince them that the machines primary language, RPG, isn't really that bad even though it suffers from 30yr old syntax H,F,D,P specs.

I think the big issue more than anything is availability of the IBM i to people wanting to know more. Let's say I am a 1M/yr company interested in the IBM i - how do I find out the truth of its features? How do I kick the tires without actually buying a machine for a minimum of $15,000 (btw, that is roughly the price for a entry level Power6 520 which I just ordered this week).

IBM has been making progress in this arena with things like their Virtual Loaner Program (only available to PartnerWorld members) and Enterprise Modernization Sandbox, but it isn't enough to keep the box from declining.

What I think IBM should do is make a trimmed down VM (virtual machine) of IBM i that can run on Intel processors so people could download it in a couple hours, install it, and have license to use it for free for up to 5 users. I don't care if I don't have an optical drive, fax line, etc - I just need ethernet and a hard drive. That is the price of entry these days. BTW, did I mention I downloaded OpenSolaris (www.opensolaris.com)the other day and had it up and running in VirtualBox (www.virtualbox.org)in less than 1hr!!?? Wonder what Oracle is going to do with that...

For the record I LOVE the IBM i and all of it's integrated features. The problem is that IBM doesn't hardly advertise those features anymore (RPG what?). Instead they advertise technology stacks that don't lock people into the machine (i.e. EGL and PHP can go anywhere). IBM needs to remember that if you do it right and the customer likes it, then proprietary is a great thing because it means you, IBM, own it all and can change it at whatever pace you want vs. risking the farm on Java which is now owned by one of your biggest professional services competitors.

The high treshhold to take just to simply start developing on a system i hasn't hurt Ibm in the days the as/400 was considered a solid unbreakable business machine, as was the case in the nineties. Meanwhile Oracle is offering its software for free for years (download www.oracle.com) and as for microsoft, although not legal, every IT student has some hacked copy of win xp and visual studio lying around, and then there is the slew of all open softwares.

But not so for system i; I don't see a student cough up $5-10K for a (2nd hand) entry level machine - provided they still want to. Why Ibm hasn't advertised and offered yet a free time share service on a system i is beyond me.

As far as proprietary softw. Every machine or software, in order to sell, needs to have a strong and distinct personality, like 'I' can do what no other machine or software can achieve. The distinct personality system i always had was the combo "command and cl language, rpg & cobol, dds, query, 5250 screens" and the suite of integrated applications from isv's. This is no longer considered valuable, although daily reality in most i shops. IBM tries to prolong life expectance by bringing open softwares like java, php, mysql to the platform. Doesn't help personality of the 'i' much either. C'mon, does java, linux, php run any better on the i then on a cheap wintel box?

How can Ibm turn the tide? If the question was posed 12 years ago, the answer was easy. But Ibm didn't act then because the money came in too easy. What to do today? Another ad in the WSJ, advertising what? Even Ibm doesn't know.

You have a great Idea about a downloadable copy of the os. We would all be using OS2 if IBM would have been able to compete with Mr Gates in the 80's.

I do not think IBM would lose $$ if they gave or sold at $1000 copy of the OS for 2-3 users and let them code Apache apps and then post to web. This would provide these people with a rock solid platform and experience. These people are not going to put down $15,000 plus software maint and hardware maint to try something!!!

Now after they used the system a while and wanted a system that could handle heavy transcation volumes, they have a option of upgrading what they have.

Now they just purchase another intel server

That's a good point, Aaron. For awhile in recent years, open source developers had their compilers and worked with them.
..At some point, Microsoft realized they had to help those developers and started giving away their limited "Express Edition" of the dot-net and Visual Studio.
..So they bought more time for their propietary development and run-time platform by giving programmers a way to do real stuff, but when they get sophisticated they must pay.
..IBM?

iManifest Japan came as a surprise - a nice one. iManifest Europe came as another jumpstart. Now it's very good to see iManifest USA keep the ball rolling. I'm feeling more optimistic.

Some of my friends work for IBM i business partners which are selling several new IBM i servers every month - usually with a proprietary software/support offering bundled in - part of the overall ecosystem. It's good to see folks like that proclaim their (and our) commitment to the platform to others outside the community.

The name iManifest says a lot, in and of itself, in terms of commitment.

COMMON's focus has shifted gradually to support Windows, AIX, and Linux, in addition to IBM i - with a primary focus on a couple annual conferences/shows.

iSociety has been useful in fomenting and operating a fan club, and other interests within the community.

iManifest focuses on the interests of business partners - who must reach out to a broader community.

The idea of IBM attracting young programmers by offering servers to play on is a nice one. But the viability and future success of the platform rests mostly on ISVs, business partners, and customers developing and deploying new applications for the platform. That takes real commitment. In most cases it takes significant infrastructure and frameworks of their own.

We need significant investment - even beyond an initial broadcast message. A manifest is a great step forward.

-Nathan

You cannot run a i5 os on an intel or any other processor.

The Power PC has a special instructions which verifies the pointer. It had to be put into the Power PC specially for the iSeries.

Part of why the iSeries is so solid. You can't mess with the pointers.

IBM i on intel? It would be nice if it ran on that chipset, but IBM would have to market the hardware. One of the key problems with windows is that it does NOT control the HW. There is a LOT of stuff in the microcode that make the i work so well and that microcode is written for the known IBM HW that i runs on. If you allow a "generic" microcode for a "generic" intel, it will break quickly. Just like windows does when it runs into a dll that was written by someone other than MS or encounters a HW mix that it doesn't handle.

Amen brother Aaron. If it is technically impossible to run a virtual on x86 then how about free hosted virtuals!?! IBM!??? While I am at it where's my href="http://news.techworld.com/applications/3202603/microsoft-offers-free-development-tools-to-websparks/?email">free
tools? http://news.techworld.com/applications/3202603/microsoft-offers-free-development-tools-to-websparks/

What does everyone think about this...

What if IBM could develop an "IBM i in the cloud" machine that supported hundreds of small LPARs, where you had complete control over your entire LPAR, and paid a small fee per month to run either 5250 apps or RPG+CGI apps? By small fee I thinking $50/mth for an entry level LPAR. Then with their capacity-on-demand capabilities you could dynamically purchase additional CPU and/or disk at certain times of the day, month, or year.

The current iSeries machine I am leasing costs about $130/month. I really have zero need to have it located in the same building I am in as I never even touch it. I would much rather lease ALL of the hardware and ALL of the software (i.e. OS, compilers, etc) virtually.

IBM has VERY cool automated virtualization capabilities in their Virtual Loaner Program that I have been using for about 4yrs now. I can basically get a fresh install of IBM i 6.1 up and running in 2hrs flat. The browser based process actually only takes me 10 minutes to setup, and the remainder of the time is their automated process carving out a preconfigured IBM i LPAR. Even better, I can save off "images" of the entire "reservation" for use later (think of it as being similar to a VMWare instance of Windows that is stored in a single file).

Seems like all they are lacking is a business model to monetize their current capabilities - they already have the technology figured out.

Thoughts?

Aaron Bartell

Congratulations on initiating iManifest in the U.S. We, iManifest Japan Board members are all excited that the initiative is now a global move.
Please let us know if there is anything you would like to know about our current activities or we can be of help.
Soon after the full page Ad. in Nikkei Economic Journal in January 2009, we started follow up actions to carry iManifest further in our market. We have set up seven task force teams to tackle various issues in promoting IBM 'i' in Japan.

S. Kakizawa
Director, iManifest Board Japan
(President, CEO Sanwa ComTec K.K. TOKYO)

Run i5/OS on Intel cpu's won't ever happen. That would destroy the last bit of pride the os and ibm has left.

ugeerts, if they can't/won't get it to run on Intel then it would be great if they would sell a small 1U rackmount appliance with IBM i installed on it for $1,500. This would allow hobbyists to easily hop on board with the machine and have their own in the basement. Access to a machine will help to address the next issue of there not being enough developed applications on the machine (applications sell the machine).

Aaron Bartell
http://mowyourlawn.com

Aaron, I love the youthfull enthousiasm in your proposal, but it is futile. Such a proposal would be good if it was part of very long term strategy to get people hooked on a machine. Something Apple has done for years in highschools; distribute free machines to IT classes. But first we are talking about servers, not personal machines, and secondly Ibm never gave this kind of free ride. The result is that no one knows anymore what system i is and worse, we're facing an aging workforce averaging 50y and going up. It frightens me to think what will happen in 10 years. Will we see a bunch of retirees working hastely on the last i machines to get them migrated to SAP?

What Ibm needs to do is to apply some kind of shock therapy to combat the messy, fckdup situation they've made up all by themselves. Will the Manifest be that shock therapy? Or are these shrinking groups of ISV vendors fighting the (last?) fight of their lives? I honestly don't know. Ibm and everyone else working in the system i arena has to draw his own conclusion.

iBM wouldn't port i to Intel because first and foremost it runs on their Power processor. I could list the disadvantages of trying to sell i/OS on Intel all day long and the marginal additional sales on Intel would not come close to justifying the risk nor replicate the iseries experience for the customer, tarnishing the iseries brand in the process.

Concerning the $15,000 iseries 520, I believe iBM is right in their comparisons to M$ equivalently equipped server. While lots of talk of freebies bandied about here, start comparing costs with a robust Intel server and start adding up the M$ multi-user production licensing for Windows Server, SQLServer, and Studio, and then start adding up third party software to attempt to duplicate i ~ i5/OS ~ OS/400 functionality.

I'd be real surprised if it comes in less than $15,000.

Having said that, Oracle buying Sun is now the competitor iBM has to face up to. With OpenSolaris open source for Intel processors and not limited to Sun Sparc processors (unlike iBM as mentioned above), its only the Oracle licensing which you have to add up, which of course is what Oracle wants.

But given Oracle's licensing for equivalent multi-user production and software development, I wouldn't be surprised if the iseries is still competitive with that setup.

Concerning aging iseries programmers, yes, I'm an old timer, but I've worked through the years with people on the iseries ~ AS/400 generally younger than me. When I interviewed throughout 2004 it was usually with younger people.

No, there are not reams of 20 somethings hired in to the iseries world like there are into Java and C#/BASIC, but when there's a business need new people are added and they are younger.

The problem is the business need. You can't outsource everything to cheap labor and still have a business need when you don't have a business.

Lastly, the single most useful thing we can do to increase interest in the iseries is to have enough useful open source ILE software (RPG, COBOL, C++, CL, etc. integrated with Java, PHP, etc.) to make buying a new or used iseries a reason to run this specialized software.

A different way of saying this is that if all people need are Java or PHP on Intel, then fine, no one needs the iseries anymore.

I know large businesses need it, obviously the less you do the less you need. But I think even for lesser needs some integrated open source software utilizing full capabilities of the iseries would do more to show the advantages of running on an iseries than anything else.

rd

Let's go a bit sideways for a moment. This kind of debate always seems to be about IBM i longevity. You know - talk of developers over 50 going extinct and such.

We have to think of a server more as an appliance, or a machine fit for a specific job. The i is like that: highly tuned as an available, high performance transaction-handling server. The language issue (usually about RPG) is irrelevant - a means to an end that developers (even the ones over 50) can take or leave on merit rather than loyalty.

I've posted on Chris's blog before as an early (actually first in the UK) i on Blade adopter. Maybe that's part of the issue - I look at my racks and see all sorts of 'best of breed' servers humming away in BladeCenters, doing different jobs according to their talents.

IBM if you're listening (hmmm ...) can you max out (sorry Chris) on diverse and specialised servers rather than looking like you're trying to converge onto a mythical single hardware and/or OS please?

Thank you. That will really help in the real world.

Ralph, imho Oracle is in no way competition to Ibm. Oracle owns, tru Sun, less than 4% of the server hardware market. A much more serious competition, somebody Ibm is really scared about and who they watch like a hawk, is HP. Both have about an equal share of 25% each. BTW, the only Ibm product that would compete against Oracle is DB2 for Wintel, which is a neglectible midget in the wintel rdbms space where Oracle and Ms SQL server share the biggest part.

But coming back to the iSeries decline & survival discussions, I notice Ibm is frequently depicted as being plain stupid for loosing business by neglecting the "i". Ibm loosing business? None is further away from the truth. Every Oracle based erp app sold has a 30% chance to run on Ibm gear, and chances are Ibm consultants will take care of installation and tuning.
Recurring income thru services, remember, and I can assure you Ibm has much more Oracle and SAP consultants running around than the few iseries guys they still have on the payroll.
The only players hurt by the decline of the "i" are the ISV's (fleeding to Microsoft Netframe based Visual Rpg) and the thoose poor guys looking for jobs on iseries. I think everybody will agree if I claim that if you compare a CV with 10 years experience in Oracle/SAP or in Iseries RPG, the Oracle/SAP guy has at least 5 to 10 times more job offers to choose from than the iseries guy. And that's what hurts with most readers here. As to Ibm, they just keep humming along to make their shareholders happy with the next quarter report.

ugeerts, you completely underestimate Oracle. They now have ability to integrate every aspect top to bottom, proprietary CPU's like IBM, Sun servers which are market leading but getting clobbered just like IBM and HP, Solaris which out Unixes AIX, ownership of Java which their software infrastructure is based on, ownership of leading apps like Peoplesoft and JDE being integrated into Project One, high end software development tools, Oracle database which is obviously #1 market leader, ownership of OpenOffice (IBM has been publicly competing with Lotus/Notes holdovers reworked with Java), high end SAN storage, everything IBM talks about and covets.

They will take the iseries market with integration of that stack if executives at IBM sit idly by and think like you do. I wouldn't put it past them given their lack of anything to go on.

rd

Ralph, you obviously got to know Oracle better lately. I've worked in an Oracle environment (pl/sql-java with web interface tru application express) few years ago and after one month already, I had to conclude how much further advanced they were and at the same time, how much Ibm left the iseries developers in the dark with some crumbs like free format rpg and integration with java, a real shame. Ibm is lucky they didn't allow Oracle on iseries and zseries (mainframe), that would have given a serious blow to db2. Oracle runs on unix-aix (pseries) and wintel (xseries).
But yes, oracle, via Peoplesoft and there own suite of integrated applications, is a threat to iseries high end customers. In Europe, the largest iseries killer is German based SAP, probably in the US, it is Oracle. The low end iseries is killed by Microsoft Sql server based erp applications, but I guess that isn't a secret to anybody.
Oracle now owns JDE, one of the last credible iseries ISV's. Oracle can close down JDE with a fingersnap - they promise not to, even released a new version of JD.Edwards erp - but the treat nevertheless stays.

Boy do I appreciate the comments from Aaron and the others on this topic. The reading has been great.

While I am impressed at the depth and width of the knowledge of the commentary and those doing the commentary, I am really concerned about what the companies who pay the $$$ for the systems and software are going to do and if the "i" is still the way for them to go.

There seems to be a trend (at least in my world) of companies going from big ERP systems, etc... to more specialized systems for different aspects of the business. These systems interact via SOA etc... Not all of them are System i packages. Microsoft, SAP, and Oracle have their fangs poised to dig into the System i ERP territory.

Companies want to get more agile, crunch data faster, dash boards, web enabled, etc... etc... etc... with less cost, less time spend developing, and in Microsoft's catch line "Can do IT".

I love the i. It has been very good to me, but I can't always defend it anymore.

We need stuff to take to the C level executives that cut the checks, make the decisions, who are getting hit with advertising for the rest of the computing world during golf, football, and baseball on TV. How does making a smarter planet have to do with running a business on the lean with less time developing solutions and more time spent implementing them?

I was amazed at how fast I got a WAMP server up and running and how fast I learned enough PHP and mySQL to get some basic apps written. That is the speed that is needed in IT. And I downloaded most of it for free to get it going. I am not an expert but hey I learned something new and it was fun.

I wish RPG was that portable. More people would use it because the language is so powerful and fast to code and get results. The snickers I get from other developers when I say I code in RPG does get old after awhile. I know the power and usefulness of modern RPG, I wish others in IT did as well.

I hope IBM is listening to us "insiders" as we hash out these topics and figure out where our future is going to lie.

- Mike Moegling