In-Depth

CICS: The Old Workhorse Keeps on Kicking

Big Blue’s durable workhorse still has plenty of kick left in it. And customers, for that matter, are increasingly willing to deploy it in support of new workloads

• By Stephen Swoyer
• 02/01/2005

IBM’s venerable CICS transaction-processing platform turned 35 last year. As it turns out, 2004 was a surprisingly eventful year for CICS—IBM even fleshed out several long-standing gaps in its suite of CICS management tools—and mainframe sales, in particular, continued to set an encouraging, if not torrid, pace. CICS appears well positioned for the future. In fact, if feedback from several long-time users is any indication, Big Blue’s durable workhorse still has plenty of kick left in it. What’s more, there are signs that customers are more willing to deploy it in support of non-traditional workloads.

One clue to CICS’ vitality came last year when IBM released a pair of new CICS-specific offerings, claiming, in the process, that by doing so, it had filled at least one long-standing marketplace gap (see (http://www.esj.com/Enterprise/article.aspx?EditorialsID=1064). Hogwash, said competitor BMC Software Corp., which argued that no such gap existed because it had provided similar tools for more than a decade (see (http://www.esj.com/Enterprise/article.aspx?EditorialsID=1091).

“We think IBM is just doing this now because they realize, because of how strong sales of mainframe software are, they know there’s money to be made here,” said Rick Weaver, a product marketing manager with BMC, at the time.

Who knew that the aging CICS could provoke such acrimony? But the fact of the matter is that IBM’s mainframe transaction-processing workhorse is as pervasive as ever. Just ask Norman Hollander, a senior mainframe consultant for Computer Associates International Inc., who’s seen firsthand just how pervasive it is.

“I have had the opportunity to see how many sites are using CICS and the related new products in their environment,” says Hollander, who has logged time with another prominent provider of CICS monitoring tools. “Some of these sites would be considered small in size; but most that I’ve visited may be considered extremely large—more than 50,000 MIPs.”

In this respect, Hollander says, CICS continues to be a very important platform for CA and other mainframe customers. Most, of course, continue to use CICS in legacy mode, but many are also retrofitting it for new roles as well.

“Customers appear to be using CICS in the legacy mode that have always done. They have been enhancing the CICS applications by integrating Java-coded applications,” Hollander confirms, noting that the number of new Java programmers in the mainframe world appears to be on the rise.

In many cases, customers are still using CICS to support applications that are decades old. Take Radoslaw Skorupka, a mainframe systems programmer with a financial services firm based in Poland. Like many other banks, Skorupka’s employer uses its mainframe systems—and CICS-based applications—to handle an important workload: All of the transactions associated with the bank’s day-to-day business. Could Skorupka imagine using any other TP other than CICS? Not a chance: “Our banking system is based on CICS. All the transactions are processed through [CICS]. We don't plan to change it.”

This isn’t surprising, says Hollander, who notes that “for old applications, CICS continues as the industrial strength application.”

Because of its reputation for stability, reliability, and scalability, CICS are also increasingly exposing CICS to heterogeneous data sources. “Database connectivity outside of CICS is still on the rise,” he points out. “Customers are using their legacy systems while crossing over to DB2 applications, IMS/DB, and CA-Datacomm.” What’s more, Hollander says, many customers are looking to bring it all back home, so to speak, to Big Iron. “SAP and Oracle on the mainframe platform is appealing to more sites.”

As a consequence of the increasing use of Java-coded applications along with traditional COBOL-based programs, WebSphere MQ implementations are spiking in mainframe environments. More importantly, customers are frequently mixing resources, tapping, for example, application server middleware (such as WebSphere and IBM’s CICS Transaction Gateway) to integrate applications running on distributed systems with CICS.

“Many sites are implementing new applications using mainframe versions of the HTTP server and WebSphere Application Server along with their current legacy CICS applications,” he comments. “Some sites exploit their legacy CICS applications from distributed versions of the HTTP servers and WAS.”

One upshot of this, for some customers, anyway, is a reduction in overall operating costs. “The mixed environments seem to help keep mainframe software costs down by offloading the MIPs needed to non-mainframe platforms,” Hollander notes, adding that Big Iron Linux has also played a role in cost cutting. “Linux for z/Series in LPARs or as zVM guests in [Integrated Facilities for Linux] are really starting to be exploited due to the attractive hardware pricing and software breaks.”

That may be the case, say Skorupka and other mainframe professionals, but Big Iron compute capacity, and CICS licensing in particular, are still prohibitively expensive. “The only real cost is the [CICS] license fee, which we consider expensive, and our staff salaries” Skorupka notes.

In fact, says Richard Pinion, a CICS programmer with a Tennessee-based health care system, mainframe support costs are negligible. “[CICS is] not expensive to support, but the software pricing can always be less.” As a result, Pinion—like a number of Big Iron veterans—is somewhat pessimistic about the mainframe’s future. “Today [CICS] is important, but I don't know if there will even be a mainframe here in five years,” he concludes.

The best way to ensure CICS viability for the future, they say, is to price it as aggressively as IBM has priced many of its mainframe Linux offerings.