HP Goes Visual with Application Modernization Tools

Published: April 15, 2008

by Timothy Prickett Morgan

The product side of Hewlett-Packard launched its umpteeth assault on the IBM mainframe and AS/400 server base a year and a half ago, and last week HP Services, which actually does server migrations on behalf of customers, announced its companion set of service offerings for legacy application migration, collectively known as the Modernization Factory. The neat bits of the Migration Factory offering are the tools that HP has created to analyze legacy applications to help customers figure out where and what to modernize.

HP is always chasing mainframe and AS/400 accounts to try to talk them into moving to Unix, Windows, or Linux platforms. But after a steady stream of application modernization efforts in the past decade, HP thinks users of proprietary machines are getting ready for big moves. "We are seeing a continuing effort in information technology and application modernization, but the momentum is changing," says Paul Evans, director of application modernization services at HP Services. "A small portion of the market acted in the past several years, but most waited. Now, a large portion of the market is on the move because they see that application modernization does work, that it does deliver TCO or ROI or whatever metrics they want to measure it by."

People do, of course, think visually, which means the code visualization tools that HP Services has created to analyze legacy applications are not just useful for the technicians who do the modernization. The eureka moment for HP has been that these tools can actually be used as part of the sales process to show customers how their legacy code is mapped out, what chunks of code are reused, how modules are integrated, and how the different code makes use of CPU, memory, and I/O resources. The latter bit helps customers and HP technicians visualize and categorize the complexity of bits of code and the and it also helps them figure out what code can be rehosted, perhaps on a Windows, Linux, or Unix platform using one of several rehosting environments and, equally importantly, what code should probably be left alone for now because it is tied very tightly to the mainframe or AS/400 to System i server.

Take a look at this image, which was created from a mainframe application set using a tool called Miner3D:

Call it eye candy for nerds, right?

According to Evans, a lot of the applications running on proprietary midrange and mainframe machinery are doing extract and loading of data, transforming it from one application to another and sometimes from one architecture to another. In this stack of applications, the ones in Quadrant 4 shown above have a lower payback in terms of modernization; they are not dependent on each other, they do not hammer I/O subsystems particularly, and they do not take a lot of resources. The applications in Quadrant 1 are probably, according to Evans, a System z10 mainframe upgrade in the making. The applications are complex, their very high I/O, and as it turns out, modernization efforts (if successful) provide the greatest payback.

Another tool, called the Clone Set Analyzer, burns through the code running in legacy systems and finds the frequency of code reuse in the applications. Take a look:

As you can see, this tool not only shows the size of reused code bits, but their frequency. That means you can figure out how to streamline the code--getting rid of the reuse and perhaps modularizing the code with the help of third-party tools, so you only have one instance of a bit of code--which means maintaining the code is simpler going forward.

While the Clone Set Analyzer shows code reuse, the Clone Pattern Analyzer tool shows how code modules are linked together using the kind of social networking analysis tools that the National Security Agency used in the aftermath of the 9/11 terrorist attacks to find out who was linked to whom and how. Check it out:

The Clone Pattern Analyzer is not a static image, by the way, but a 3D set of representations of linked code modules that you can drill down into much like you can on the Web with hypertext documents. By knowing how modules are linked together, HP and its application modernization customers can figure out where to start modernizing and see immediately what other modules will be immediately affected. "As soon as you make this visual for customers, the first thing they say to me is, 'My code cannot be that bad,'" says Evans. "But then they realize that it is."

HP is making these new code analysis tools in educational centers it is calling Modernization Factories, which help rehost applications on emulation environments or help them decompose their applications and re-engineered on new platforms with modern application development tools. The visualization tools are not sold separately by HP, but are part of a two-week HP Services engagement that costs $50,000. Evans says that HP is working on another services engagement for price-sensitive and smaller midrange shops that will cost $20,000. The tools can be used on just about any application set you can imagine, including COBOL, CICS, PL/1, and Adabas applications running on MVS-OS/390-z/OS mainframes, RPG and COBOL applications on OS/400-i5/OS machines, and even applications running on VAXen and HP 3000 gear.

RELATED STORIES

The HP Pitch on Rehosting i5/OS Applications on Integrity

TFH Flashback: Assault, Battery Not Included

Infinite Software Partners with HP, Acquires Altos Technology Group

HP, Intel, and Oracle Gang Up on IBM Mainframes

Bang for the Buck: Big Iron Boxes, Even Bigger Bucks

A Joint Assault on the Mainframe Hardware Market

Intel Aims Dual-Core Itaniums at RISC, Mainframe Servers

HP Gears Up for Montecito Itanium Shipments

California Software, Unisys Chase OS/400 Base

                     Post this story to del.icio.us
               Post this story to Digg
    Post this story to Slashdot