In-Depth
The Changing Face of Host Integration
Rather than moving host applications off the mainframe to save money, analysts suggest that it's time to look into service-oriented architecture
If you’re still using the same old terminal emulation clients or Web-to-host front-ends to expose your host applications, you're not up to date, industry watchers suggest. If the mainframe is here to stay, it’s time to think about exposing your host applications in a more intuitive and flexible manner—as part of a service-oriented architecture (SOA).
Happily, a spate of host integration vendors, along with IBM Corp. and Microsoft Corp., are eager to oblige you.
One reason for the mainframe’s resurgence, says Phil Murphy, a principal analyst with consultancy Forrester Research, is that organizations have discovered that moving host applications to Windows, Unix, or other platforms just isn’t cost effective. “We went through the 90’s saying, ‘Get off the mainframe!’ and so a bunch of people went to Unix. They saved money on hardware and software, but they spent that money on Unix administrators, and there simply weren’t huge savings,” Murphy comments. “If you’ve got under 1,000 MIPS, you can probably still move and still gain some efficiencies, but other than that, the moving for the sake of saving money is a bad deal. It doesn’t work.”
Instead, a kind of curious détente has settled over many IT organizations: For the first time in more than a decade, the mainframe isn’t under siege from without—and host integration vendors sound downright optimistic about the future. “The mentality now seems to be ‘get the best from your existing assets,’ so there’s no longer the debate about replacing host systems with Unix, or client-server, or whatever,” says Peter Havart-Simkin, senior vice-president of strategic development with host integration specialist NetManage Inc.
Consider the case of Microsoft’s Host Integration Server (HIS) 2000, which followed its erstwhile SNA Server product. Since introducing HIS more than four years ago, Microsoft has done just about everything in its power to encourage that product to go away—even killing the HIS marketing budget and slashing its development team—and yet it keeps selling.
Never one to ignore a market trend or a revenue opportunity, Microsoft more or less gave in last year, announcing at its Tech Ed 2003 conference plans to ship an updated version—HIS 2004—sometime this year. “That's a great little product for us. It isn't exactly the most strategic thing on every customer's mind, but tactically it's very important. Customers who have mainframes deployed depend [on this product]," said Paul Flessner, Microsoft senior vice president for Windows Server Systems, in an interview at Microsoft's Tech Ed show last year.
From Host Access to Host Integration
As a result, argues Eric Varness, director of product management with host integration specialist WRQ, host access and host integration technologies have assumed new importance. “There’s starting to emerge this concept of Web-to-host or legacy rejuvenation where the customer wants to extend this [mainframe data and applications] to non-traditional users, and they’re realizing that character-based [terminal emulation] just isn’t going to work,” he comments.
The idea, Varness and other industry watchers say, is to expose legacy data and applications in the context of a more flexible SOA. To that end, many of the vendors that play in this space have introduced bona fide host integration platforms—complete with support for both the J2EE and .NET application environments—designed to complement their existing terminal emulation and Web-to-host products. “Customers can extract [a mainframe application] into a series of reusable services that can then be exploited by a developer who’s building a front-end application, usually in a variety of different architectures, usually in Microsoft or [J2EE],” he explains.
In addition to its bread-and-butter host access offerings, Attachmate Corp. markets myEXTRA!, which it bills as an all-in-one host access, data integration, and presentation platform. Ditto for NetManage, which markets OnWeb, a complement to the RUMBA terminal emulation and host access technologies it acquired from the former Wall Data Inc. nearly five years ago. WRQ, for its part, markets Verastream, along with its conventional terminal emulation products.
To be sure, industry watchers acknowledge, the vast majority of customers are still using tried-and-true approaches to host access. “A lot of it is Web-to-host, as opposed to direct green screen, but basically, you’re still intercepting the green screen,” comments Mark Vanston, a program director with consultancy META Group.
The irony is that host applications are probably better suited for exposure as part of an SOA than many applications based on more modern 4GL object-oriented languages, says Forrester’s Murphy. “When folks wrote screen-based transactions many months ago, they wrote it at a business function viewpoint: I add a customer, I add an order for that customer, I check backlogs for that customer, etc. So in many respects, those CICS screens of 15 years ago are better suited to service orientation than a lot of the newer, distributed code that’s been written over the last several years, because of their affinity with a business function,” he argues, adding: “What did the object-oriented guys do? They took those screens and they broke them down into a thousand different objects.”
Because IBM has such an extensive portfolio of terminal emulation, Web-to-host, and host integration offerings—most of which are heavily dependent on its WebSphere J2EE application server—NetManage, WRQ, and others stress the .NET friendliness of their own offerings. “The other valuable lesson that’s come out is that most users are only looking for one development environment to do all of this, so if you’re a Microsoft shop, you’ll want to do everything in Visual Studio,” comments Havart-Simkin, who stresses that NetManage’s products also support J2EE. “The big difference is that the sort of server technology that we have running on Windows is able to do the equivalent of all of those different IBM platforms running on WebSphere.”
In addition, say Havart-Simkins and other host integration proponents, because .NET is a mobile-friendly application architecture, it’s actually a better choice for organizations that want to expose host applications and data to PDAs and other wireless clients. “This is where Microsoft will score, because the .NET framework goes right down to those devices, through the Compact .NET framework, so it’s very, very easy to develop for them,” he asserts.
Finally, host integration vendors argue, their products typically support host platforms other than IBM’s mainframe or iSeries midrange servers, such as the HP 3000 and OpenVMS platforms from Hewlett-Packard Co. (HP). “There are still people out there with the HP 3000 and OpenVMS, especially, which is actually a very active community. You’d be surprised how religious those people are about that platform,” says WRQ’s Varness, citing estimates from HP that put the number of extant OpenVMS customers at 25,000 or more.
SOA Isn't the Answer for Everyone
Not everyone is persuaded that SOAs involving mainframes and other host systems are inevitable, however—at least in the near-term. “I think everyone’s looking at it, some are building services around it, and there’s a lot of XML translation going on, but I’m not sure if the whole Web services model has been fleshed out enough that people are comfortable with it,” comments META Group’s Vanston. “But is the application layer really changing? No --- you’re basically still doing some kind of interceptor strategy.”
Indeed, Microsoft’s decision to revamp HIS came in spite of its best efforts to establish a more sophisticated offering, BizTalk Server, which it released in late 2000, as a Web-services-enabled platform for enterprise application integration.
When the software giant introduced BizTalk Server, which supports XML and uses sophisticated adapters to facilitate access to mainframe data and applications, it quite naturally expected that its HIS revenues would take a beating. That hasn’t been the case, however. “I thought they'd dip. But they kept going along. We don't market it. I took the development team down to a very small team. From a peak of about 80 people down to about 6. But it's still a very critical product to customers who want it," Flessner acknowledged last year.
Although host integration proponents acknowledge that only a small percentage of customers are today building SOAs that involve their mainframe systems, they argue that as IT spending picks up, SOA investments will, too.
“With the invigoration of IT spending, people are going to go back and start to do more of these, updating their business processes on the front-end so that they can get that 360 degree view,” suggests WRQ’s Varness. “That CRM initiative that stalled for the last three years is going to start again, and if you’re going to get that 360 degree view, guess what? You’re going to have to get a lot of that [data] from the mainframe.”