CA talk on LISA platform well attended by govt IT
Subscribe now for $100 (23 issues) and save more than 37% off the cover price!
Get the latest news from Computerworld delivered via email.
Sign up now
The IT world is galloping toward cloud computing and software as a service and service-oriented architecture but is struggling to deal with ever-increasing complexity, according to CA.
Components are increasingly interconnected; development and other teams become interdependent, and heterogeneous technologies have to talk to each other, the company says.
As far back as 2008, IDC reported that the testing backlog was the single largest factor in the delay of new application developments. The same year, Gartner said that “unplanned downtime” increased by 50 percent for SOA-based, loosely couple applications.
Some other observations presented at a recent, well-attended (mainly by government departments) seminar by CA to talk about its LISA platform for composite agile development highlighted several issues including:
• Eighty percent of all new applications are composite and highly integrated;
• Most new application development is self-service and customer facing, increasing complexity;
• There is limited predictability on how applications will perform and when they’re likely to break.
The premise behind CA’s LISA platform is to virtualise everything to eliminate constraints when modelling complex applications.
Jim Starr, CA’s virtualisation practice director, Asia-Pacific and Japan, was for seven years the general manager of quality and testing at Telstra. He ran 50-plus test environments, managed five global releases a year, up to six other releases in any given month, and dealt with multiple architectures.
“It was very costly and very manual,” he says. “We wanted automation and something that was highly virtualised. We identified LISA as best of breed.”
LISA was developed by US company iTKO, subsequently bought by CA.
“We automated 25 applications, such as portals,” Starr says. “More than 30 downstream services, such as Amdocs, were virtualised. We reduced global regression testing from two weeks to three days.
“Virtualisation is changing the way software is built.”
Telstra was able to take the specifications for Australia’s National Broadband Network and create test case capabilities to join the network six months before anything was built. “We could simulate close to the real thing,” Starr says.
CA claims LISA produces efficiency gains of up to 30 percent, and finds up to 80 percent more defects prior to user acceptance. That produces big savings by avoiding infrastructure upgrade, the company says. The seminar was hosted by CA business partner Activate New Zealand.
Unfortunately the LISA platform will not reduce the complexity of IT projects unless the business complexity is minimized up front and IT architecture reflects business architecture rather than business having to mirror IT.
Delivering software projects 30% quicker only means you have to address the failures 30% quicker. And it still does not mean that the application will do what business needed it to do, nor improve the quality of requirements or eliminate scope-creep, because you still have the same poor quality information, wrong people involved, lack of planning and very slack governance measures.
The LISA platform cannot fix what is broken across the business aspects of commercial enterprise or government entities undertaking IT projects. These are the critical and most difficult issues that need to be addressed - people, visibility, communications, quality executive decision making, business planning and governance.
When you have people and government bureaucrats believing that there is a silver bullet for their IT project failure woes, they will invest so much money but still wind up back at the same point they are in now. Albeit a few millions lighter in the pocket.
Additionally, they may be able to detect 80% more defects, but does the software do what business needs it to do? Probably not, because if you look at any government IT project in NZ or Australia, you will find that the project is either scrapped, doesn't deliver or massive cost and time over-runs.
Finding defects is not the answer! Getting business and executives to take accountability for their role in planning, governance and accurate requirements before the project is invested in or commenced is the only solution. Then, and only then can you even think about finding defects or increasing efficiency.
Lets check back in about 6-12 months, but I would put my money on me.
Posted by Sarah Jane Runge at 18:29:39 on October 5, 2012