Securing top-down support key to Blue Cross mainframe consolidation |
 |
By Mark Fontecchio, News Writer
04 Mar 2009 | SearchDataCenter.com |
|
|
This week, Blue Cross and Blue Shield of South Carolina received an award for technology excellence from the Share mainframe user group, mainly for the organization's consolidation of distributed servers onto a mainframe that runs about 250 virtual Linux server instances.
SearchDataCenter.com discussed the project with Bhanu Rai, Blue Cross' zLinux IT manager, and how his company got it to work.
How did this project come about? When did it start?
Bhanu Rai: It started around July or August of 2007. The initial effort for the zLinux decision was made primarily by the CIO and the chief technology architect. It started as a refresh project. We had the opportunity to buy a bunch of new [distributed] servers. Since we're already a big mainframe shop, we decided that instead of doing that, we'd try doing it on zLinux. I came into the picture in October 2007 to lead the team to complete this project.
Why did you choose the mainframe over distributed servers?
Rai: We know our mainframe best practices have worked for 20 years, and we have a stable architecture around the mainframe. We thought we could make it work like the rest of the mainframe. Virtualization has been on the mainframe for many years. We run a bunch of CICS instances on one operating system, for example.
|
|
|
|
|
There needs to be a top-down approach.
Bhanu Rai,
zLinux IT manager,, Blue Cross and Blue Shield of South Carolina
|
|
|
|
|
|
|
|
|
|
We took a look at Nationwide Insurance [and its zLinux project], and based on all the feedback we got, we decided to go with the mainframe. Of course they did [a return on investment] study as well. I can't say how much they saved exactly, but it's all about the money. If they weren't going to save money, they wouldn't do it.
Was it difficult to sell the project internally?
Rai: I was in technical work and was always in the background, but it was a big thing, because there needs to be a top-down approach. Some people are not open to change, and we needed to get everyone on board to give it a shot. It was a big deal getting the executives not only supporting the project but leading it.
Getting support from up top at the start was very important, and they were pushing the effort, too.
So what applications did you move from distributed to the mainframe?
Rai: Customer service desktop applications. It was a front-end, customer-facing [graphical user interface] app, a Java app running on x86 that is now running on WebSphere on zLinux. The front end still looks the same to employees. It's just running on the mainframe now.
Did you run into any hurdles?
Rai:One thing we started finding in these applications were default best practices for Java. That was stopping us from consolidating to one Linux server. It was kind of like if you go to the CNN website and you can choose national news or international news. Because of best practices, it was like you could choose only national news. So we had to change some of the applications a bit so they would fit better together.
What's in store for the future?
Rai: We have other applications lined up to move to zLinux. Not everything is a good candidate. The chief technology architect decides which applications give us the most bang for the buck, and we do them first.
What lessons have you learned?
Rai: When we started, we had to tell everybody to start with a different mindset. If you think of it as a distributed server project, and you have an x86 Linux mess, you could end up with a zLinux mess. Fortunately, we have a good system architecture guiding the development process.
For more on the Share conference, check out our 2009 Share conference coverage page.
Let us know what you think about the story; email Mark Fontecchio, News Writer. You can also check out our Mainframe blog.
|
|
|
|
