System iNetwork

November 2007

November 27, 2007 1:19 PM

Evidence Grows That App Modernization Tops App Replacement

The assertion that modernizing your applications is a better investment than replacing them outright recently gained a little more credibility in the mainframe world. I suspect that's likely true for at least some parts of the System i market as well.

A recent study by Forrester Consulting conducted for Software AG, maker of Adabas (a data-management solution) and Natural (an application and deployment environment), both mainframe tools, found that legacy modernization would produce a five-year return on investment (ROI) of 331 percent, with payback achieved in less than four months in a composite case study of several representative enterprises. Titled "The Total Economic Impact of Maintaining Adabas and Natural," the study went on to report other findings about Adabas and Natural specifically that don't concern us but also noted a few other things that at least seem to apply to the System i.

I'll be the first to admit that the mainframe world and the System i market only touch at certain points and that it may be overdoing it to read too much into studies and surveys about Big Iron shops. However, at the same time, looking over the history of just the past 10 years in our market shows an accelerating convergence between these two universes that were once starkly separate. The two platforms lately have come to have a number of similarities. High-end System i models have mainframe horsepower, and enterprises that were once considered too large to be part of the System i market are finding it just as economical, if not more so, to operate a cluster of i5s instead of one or two System z-type machines. So the platforms' markets are overlapping more than before, a fact acknowledged by IBM's own recent reorganization to key its sales efforts to potential customer size rather than specific platforms.

What's more, there's a mutual theme of "relevancy" that's plaguing both the System i and its larger counterparts. Large businesses have been tinkering with applications running on the System z for years, and smaller businesses have been maintaining applications running on the System i for a similar period. Now both kinds of enterprises are starting to face pressure to consider whether "the way we've always done it" is "the way we should continue to operate," and application modernization is one of the touchstones of this debate. (As it should be, really. What workers don't have to pause now and again to sharpen their tools? Information workers are no different.)

Enterprises interviewed for the study include a European branch of an automobile manufacturer, a U.S. health-insurance provider, an agency of the Australian government, and a France-based financial-services firm. Such organizations could easily be System i rather than System z shops. Participants were considering modernization because they feared their existing systems were becoming outmoded.

In addition to the hearty endorsement of "update and extend" application-development strategies as opposed to the "rip and replace" idea for most organizations, the study found that many users at these enterprises were basically happy with such elements of current systems as transaction speed, reliability, and ease of use (at least, with regards to their personal familiarity with using the software). Most users would be satisfied simply with improvements to UIs (such as browser accessibility) and better integration between existing applications. They were wary of risks involved with having to deal with learning to navigate new systems that probably wouldn't have all the bugs worked out of them at first.

This sounds pretty familiar, doesn't it? You probably have a fair amount of anecdotal evidence that many of your end users have similar thoughts.

I wonder if, in some regards, the modernization debate isn't just another expression of a philosophical divide, namely the difference between the idealists and the pragmatists, that can make it hard for us to come to terms with other aspects of society. Some folks at the top who don't necessarily use every day all the applications that make their businesses run are sometimes the first to react to the emotional appeal of the "let's start over with something new" approach, while many of the worker bees are thinking, "if the software I have to use to get work done just did this and this as well as what it already does, it'd be good enough."

At some companies, application modernization means there are advocates for tossing the System i onto the scrap heap and opting for reengineering the business on a new platform.
Perhaps now that the modernization debate has gone on a while, this study is more than just an interesting factoid. It's maybe a clue that conventional IT wisdom has already started to turn against the "rip and replace" philosophy. If that's so, rather than a threat to the System i, isn't legacy app modernization really an opportunity to maintain the platform?

Common sense tells us that evolution is generally easier to achieve than revolution. There are plenty of packaged tools that can help System i developers with little experience in modernization jump-start the process. Factor in the expenses of creating or buying new systems and software, training people to use them, and living through the other hassles of adapting people and processes to them, and it's perhaps a bit less surprising that the Forrester study found such a dramatically better ROI for legacy app modernization over replacement. Those are also powerful arguments that it's usually more cost effective to make the best of what you have rather than trying to reinvent the wheel, and that idea favors keeping System i machines in System i shops.

Posted by on November 27, 2007 at 1:19 PM | Comments (0)

November 13, 2007 1:26 PM

System i BI Vendors View IBM/Cognos Deal Skeptically

On Monday IBM unveiled a $5 billion deal to purchase business-intelligence (BI) tool vendor Cognos. In some regards, this action can be viewed as a bit of "keeping up with the Joneses," coming in the shadow of such other BI-company consolidation moves as Oracle Corporation's $3.3 billion buy of Hyperion last spring and SAP AG's $6.8 billion hoovering of Business Objects last month. It's part of a more general recognition in the wider IT market that BI applications, already important, are going to become a crucial tool for business efficiency in the years ahead and therefore an even more significant source of revenue.

One issue in BI that's coming to the forefront and that's definitely on IBM's mind is a concept currently called "a single version of the truth." Although tying large amounts of data together better than ever before is generally a good idea, it isn't automatically useful if some of the data is faulty. Old data, human error, communications mistakes, and scores of other reasons can cause data about specific customers, products, and terms or other elements of business deals to disagree with each other. We all know where "garbage in" traditionally leads. So some attention must be paid to not simply synchronizing data but to making sure all the pieces of the information pie are correct.

What's more, you can find useful information in many places besides databases, for example in "unstructured" sources such as e-mail messages, images, custom applications, and so forth. Drawing all these threads together to make a useful collection of facts for analysis is a challenge that grows quickly with an organization's size.

IBM's plan for dealing with these problems is Information on Demand, its stated reason for emptying its deep pockets a bit for Cognos, which IBM in fact cites as its 23rd acquisition in support of Information on Demand. Information on Demand is a plan for "freeing [information] from silos and delivering it" across an enterprise. So, for example, IBM cites an ideal situation in which a call-center employee not only has current contact information on a prospective customer but also has instant access to such information as balances, transaction histories, recent e-mails, and other information to "identify up-sell and cross-sell opportunities." Well, who wouldn't want that?

IBM's plan with Information on Demand is to provide it by blending (if not hammering into a squared circle) products and their elements from all its acquisitions and ultimately offering it within a framework that supports a service-oriented architecture. Cognos's appeal, according to IBM, is that its products fit into this strategy. (It has nothing to do, of course, with the fact that Cognos is the largest independent BI vendor left alive in the general IT market.)

What's of most interest to you and me is what this means to the System i market. To begin with, Cognos products don't run natively on the System i. However, they use an open architecture, and they interact with DB2 databases on System i servers, and in that sense they support the System i and applications that run in environments that include the System i. (Cognos also participated in a BI-tool-buyer's guide, which System iNEWS will publish with its December issue.) I think it's not too fantastic to imagine that somewhere down the line, once Information on Demand is a going concern, it will be extended to the System i (although perhaps it will actually be the System i itself that ends up being extended into a form compatible with Information on Demand). Long range, I think the Cognos acquisition has an excellent chance of putting IBM on a path to providing its corporate customers with some killer BI tools (as long as Information on Demand doesn't get as gigantic as WebSphere is, and there's clearly some danger of that). The question is whether those integrated offerings will release before Vista 2020 (codenamed "Hindsight," remember you heard that here first) does.

I asked Scott Steinacher, a technical editor for System iNEWS who specializes in the BI area, for his opinion of the acquisition. "Financially, I find this deal difficult to justify. What's more, it could not have come at a worse time for IBM Rochester. It complicates marketing the new DB2 Web Query product, which is itself a port of a competing product
from Information Builders," he said.

To get a feeling for the short-term impact of the acquisition on our market, I decided to do an informal opinion poll of some System i BI application vendors. Although the responses aren't unanimous, most System i BI vendors also view DB2 Web Query as a significant consideration, though many seem skeptical of the Cognos purchase causing a large impact on the platform.

"The consolidation of companies that's underway in the BI market strengthens those companies' stacks," ventured Datawatch Corporation COO (and soon to be CEO) Ken Bero. "Among customers, though, I believe it will cause some consternation. It throws things up in the air. No one knows how to position their [BI] strategies. It makes it more difficult to make decisions. Here at Datawatch we view this as an opportunity to continue to provide easy-to-use BI tools with a strong ROI, today." As for IBM's justification for buying Cognos particularly, "it's not clear what they are thinking," Bero opined.

"I think the most significant thing about [this acquisition and SAP's of Business Objects] is that it has brought attention to the value of BI," states New Generation Software's Marketing Director, Bill Langston. "We think this will make more executives ask their IT directors and CIOs to investigate BI, and that's good for everyone in the BI market," he continues. But, he noted, Cognos products don't run on i5/OS "and I would be shocked if IBM elected to make the investment required to change that. From a technology perspective, the Cognos acquisition really doesn't change anything for System i customers."

John Hughes, Silvon Software's senior vice president, had an outlook similar to Langston's. "We don't see any significant impact to Silvon resulting from this merger since for many years we have been a System i 'application-centric' BI provider. That means our domain expertise across supply chain, manufacturing, and distribution has driven demand for our solutions."

BCD International's Eric Figura agreed with Datawatch's Bero. "While this could be a good fit for IBM, are they not simply creating a lot of marketplace confusion for System i users? Isn't IBM [already] doing a full-court press pushing Cognos's competitor Information Builders' web query tool? What is [IBM's] clear message with this news?"

A representative of another vendor company who declined to be named looked at the situation even more darkly. "[I think this means IBM's DB2 Web Query] will be superceded by this new acquisition. All the hype over [that] was for naught and will mislead many customers. It's just another sign of the disconnect between IBM Rochester and IBM New York . . .a common theme at IBM."

Doug Mack, IBM's System i DB2 product marketing manager, took strong issue with that point of view. "That is an erroneous assertion. We are already working on the next version of DB2 Web Query, and this acquisition will have no impact on that effort. For those customers looking for an i5/OS-centric solution, we expect DB2 Web Query for System i to continue to be an attractive query and reporting solution."

Information Builders, the originator of DB2 Web Query, agreed with Mack. "The Cognos acquisition will have no effect on the System i market or our ability to have a relationship with IBM. [IBM] turned to the leader in operational BI that runs natively on all platforms . . .to revive and grow its System i base," stated a spokesperson.

DB2 Web Query is also a basis of Mrc President Joe Stangarone's optimistic view. "I don't believe the Cognos acquisition was done with DB2 or the System i as a primary consideration. If it was, how could you rationalize the rollout of DB2 Web Query for the System i just a few months ago?" he asks. "The System i BI market will find [itself] shifting to smaller, more capable vendors, who are nimble and far less constrained than the behemoths. Small and mid-size vendors in the BI marketplace who can listen to their customers and deliver customized BI solutions without all the headaches will be able to offer solutions in a much more timely manner."

Posted by on November 13, 2007 at 1:26 PM | Comments (3)

November 6, 2007 9:40 AM

Web Apps Make Every Enterprise a Software Company

I'm sure we're all somewhat familiar with the software release process. Even if you're a member of the rare enterprise that uses all custom-built software applications, keeping up with the various iterations of i5/OS (V6R1? Where has my life gone!?) and Windows has attuned most of us to the cyclical nature of new versions of software and the idea that what follows has to be better than what you already have in place. But if you're moving into the brave new world of providing your Internet customers with web-based applications, even if it's just a simple app for ordering your products, have you realized how much that activity is turning your enterprise into a defacto software company?

Any enterprise that writes custom programs for its internal processes already faces this to some degree. Issues such as tracking progress on new versions, making sure software is as bug-free as possible before rolling it out to internal users, making impromptu problem fixes, and managing different releases of the same application aren't new.

But of course, what is different about web apps is the audience. Instead of all your potential users being employees of the same company, who are going to be required by their management to use a piece of software no matter how lame or cool it might be, web apps serve people who are under no compulsion to tolerate bad software. That raises the stakes and puts a real premium on producing applications that are understandable without any training and that actually do everything they're supposed to do, correctly, in version one. And that places web app development in the same league as producing software products for sale. A web app has to work well because no memo from the company president can make all the users put up with something that doesn't. The web app also becomes part of your public face, just as if you were selling it to the public. The only part you get out of is having to haggle over the price. And you don't get the benefit of paying customers pointing out your problems, either, like they do for the software product vendors. The disgruntled will simply go away.

If you're putting custom software programs onto your web site, you're basically introducing yourself to many of the same dilemmas software companies face as part of their basic business challenge. No matter how simple it is, that app is going to have to be changed some day because no process is totally static. What's even more urgent, though, is that web consumers' expectations are constantly going up as they encounter other web sites, in other business realms, that are even slicker than yours is.

Once you step into the web app arena, you're not just vying with the same competitors who've been on your radar for years, in a way you're competing with every other web site in the world. And while you might be able to hide from this problem long enough to fool an upper management that isn't sufficiently forward-thinking to see the danger of not fully participating in web commerce, you can't outrun this reality forever, and you'll be even farther behind than you are now if you don't get started really soon.

Dealing with this situation may require some thinking along different lines than some of us are used to. Concerns such as providing a GUI that's intuitive and inviting but not too cluttered, making sure there are no significant bugs even in minor functions, and always thinking ahead to what's going to be in the next version aren't optional anymore. Software company executives seem to be born thinking such thoughts, that's how they stay in business. With web apps, the rest of us have to start thinking that way, too.

The saving grace is that everyone other than the software executives are in the same bind. And how are your opposite numbers at other companies handling this new challenge? Not too well, it seems. Or at least, that's the picture presented by a recent survey released by Interwoven, a content-management solutions provider.

Interwoven conducted the survey of 130 IT operations professionals worldwide. Admittedly, that number may not be statistically significant, but even if the percentage levels might not be totally representative, I think the trends the results show have at least some validity. Interwoven's motivation in conducting the study was to promote its Custom Application Provisioning (CAP) product, which helps IT departments "automate delivery of code, content, and configuration" to speed up delivery of web apps. (CAP doesn't run on System i but does support System p, by the way.)

In summary, Interwoven's results showed 77 percent of respondents report development efforts are on custom projects, more than half said they thought their number of web-related custom app projects would double in the next year, and 71 percent are accomplishing upgrades with manual processes such as FTP, build scripts, mailing CDs, or copying files. (Interwoven refers to this as the "push and pray" method, a term that seems to be adopted from poker, in the sense of pushing something out there and praying that it works.) 73 percent said they can only accomplish one or two new web app releases a month and 60 percent feel pressured to issue releases more frequently. Also, 54 percent say they have to make up to 10 "hot fixes" a month.

Web development on the System i generally isn't advancing at this frantic a pace yet in most shops, but the need for it doesn't spare our platform. When those needs arise, though, the intensity will be as great or greater than the pressures that already accompany internal custom software development. The issues raised by this survey should be food for thought for any software development manager who sees a requirement for a more dynamic web presence for his or her enterprise coming down the pipeline. What seems clear for at least some of us is that paying more attention to the mechanics of producing web applications may be an inevitable part of our 2008 landscape. And looking to vendors of general software products for some inspiration isn't a misplaced notion.

Posted by on November 6, 2007 at 9:40 AM | Comments (0)