Friday, December 03, 2010

Case Study: AIG Insurance Group Leverages ALM to Attain IT Performance Architecture Advantage

Transcript of a sponsored BriefingsDirect podcast, part of a series on application lifecycle management and HP ALM 11 from the HP Software Universe 2010 conference in Barcelona.

Listen to the podcast. Find it on iTunes/iPod and Podcast.com. Download the transcript. Sponsor: HP.

Dana Gardner: Hello, and welcome to a special BriefingsDirect podcast series coming to you from the HP Software Universe 2010 Conference in Barcelona.

We're here in the week of November 29, 2010 to explore some major enterprise software and solutions, trends and innovations making news across HP’s ecosystem of customers, partners, and developers. [See more on HP's new ALM 11 offerings.]

I'm Dana Gardner, Principal Analyst at Interarbor Solutions, and I'll be your host throughout this series of HP sponsored Software Universe Live Discussions. [Disclosure: HP is a sponsor of BriefingsDirect podcasts.]

Our customer case study today focuses on AIG-Chartis insurance and how their business has benefited from ongoing application transformation and modernization projects.

To learn more about AIG-Chartis insurance’s innovative use of IT consolidation and application lifecycle management (ALM) best practices, please join me in welcoming Abe Naguib, Director of Global Performance Architecture and Infrastructure Engineering at AIG-Chartis in Jersey City, NJ. Welcome to the show, Abe.

Abe Naguib: Hello, Dana. Nice to be here.

Gardner: Abe, tell me a little bit about the scope of your organization and the type of applications activities that you are undergoing and why moving toward some newer products made sense?

Naguib: Let me step back for a second, Dana, and give you a background on AIG and its applications. AIG is a global insurance firm, supporting worldwide international insurance of different varieties.

We're structured with 1,500 companies and roughly about eight lines of businesses that manage those companies. Each group has their own CIO, CTO, COO structure, and I report to the global CTO.

What we look at is supporting their global architecture and performance behavioristics, if you will. One of the key things is how to federate the enterprise in terms of architecture and performance, so that we can standardize the swing over into the Java world, as well as middleware and economy of scale.

Gardner: Given the breadth and depth of the organization, where are you in terms of your applications? What are some of your goals in terms of improving how things are done?

Proliferation of middleware

Naguib: I started about 10 years back, when I came on board to standardize architecture, and I saw there was a proliferation of various middleware technologies. As we started going along, we thought about how to standardize that architecture.

As we faced more and more applications coming into the Java middleware world, we found that there’s a lot of footprint waste and there’s a lot of delivery cycles that are also slipped and wasted. So, we saw a need to control it.

After we started the architectural world, we also started the production support world and a facility for testing these environments. We started realizing, again, there were things that impacted business service level agreements (SLAs), economy of scale, even branding. So, we asked, how do we put it together?

One of the key things is, as we started the organizational performance, we were part of QA, but then we realized that we had to change our business strategy, and we thought about how to do that. One key thing is we changed our mindset from a performance testing practice to a performance engineering practice, and we've evolved now to performance architecture.

The engineering practice was focused on testing and analyzing, providing some kind of metrics. But, the performance architectural world now has influence into strategies, design practices, and resolution issues. We're currently a one-man or one-army team, kind of a paratrooper level. We're multi-skilled, from architecture, to performance, to support, and we drive resolution in the organization.

Gardner: What is your role in that team?

Naguib: I manage the organization in terms of deliveries. We hold internal best-practice discussions. We catch trends and metrics in our knowledge base. We influence design. We even influence vendors that come in. We partner with a lot of the products that come in.

So, we meet with IBM, HP, and Oracle products, and as we influence and capture trends, we work back with the product development teams to figure out how to first resolve internal development, as well as the product that we build on.

Gardner: And as you were making the transition to this performance architecture, what were some of the important considerations you had in terms of making that more holistic, more managed, and more comprehensive?

Naguib: One of the biggest things was the time to market. We also saw that resolution had to happen quickly and effectively. Carnegie Mellon did a study about five years ago and it said that post-live application resolution of performance issues was seven times the cost of pre-live [performance application resolution].

In other words, we realized that the faster we resolved issues, the faster to market, the faster we can address things, the less disruption to the delivery practices.

Too many people involved

In normal firefighting mode, architecture is involved, development is involved, and infrastructure is involved. What ends up happening is there are too many people involved. We're all scrambling, pointing fingers, looking at logs. So, we figured that the faster we get to resolution, the better for everyone to continue the train on the track.

We built a practice with architectural engineers and DBAs to get to issues and resolve them faster.

Gardner: So, when you've got multiple teams and then fairly large numbers of people involved with these teams, they're probably distributed as well. What’s the overall umbrella concept? What did you need to pull that all together and to give you that view into these activities to make that performance, integrity, and speed come together?

Naguib: The key thing is that we started working with the CIOs at that level, and figuring out a strategy to develop a service-level target, if you will. As we went along, we began working with the development teams to build a relationship with the architectural teams and the infrastructure teams.

We became more of a team model, building more of a peace-maker model. We regrouped the organization, so that rather than resolve and point fingers at each other, we resolved issues a lot faster.

Now, we're able to address the issue. We call it "isolate, identify, and resolve." At that point, if it’s a database issue, we work directly with the DBA. If it’s an infrastructure or architecture issue, we work directly with that group. We basically cut the cycle down in the last two or three years by about 70 percent.

A lot more CIOs have started bringing in more applications. We see a trend growth internally of roughly about 20-30 percent every year.



Gardner: And as you're increasing your goals of speed and integrity are you also able to handle more applications at once? Does this improve the volume of applications going through your pipeline?

Naguib: Absolutely. Because there is a change in our philosophy, in our strategy to focus more on business value, a lot more CIOs have started bringing in more applications. We see a trend growth internally of roughly about 20-30 percent every year.

I have a staff of nine. So, it’s a very agile, focused team, and they're very delivery-conscious. They're very business value-conscious, and we translate our data, the metrics that we capture, into business KPIs and infrastructure KPIs.

Because of that metric, the CIOs love what we do, because we make them look good with the business, which helps foster the relationship with the business, which helps them justify transformation in the future.

Gardner: Can you share with us any of those KPIs, what’s the report card that you could bring back to your superiors in a business sense? What’s the business case and rationale you can provide?

Footprint is key

Naguib: If you look at ITSM model, Service Level Delivery, one of the key things is the footprint of applications. One thing that organizations are starting to realize now is that software drives the hardware. For example, the cost of IBM WebSphere on hardware is much more expensive than actually buying a server.

In traditional firefighting mode, people tend to hire consultants, bring in hardware, and end up increasing their cost. What we found is that, if you address the software angle of it, then you can improve your TCO and ROI.

By taking a looking at correlating business transactions to a footprint on a server, and improving those transactions and their consumption rate, you're actually effectively improving the consumption of that application particularly. And as you improve that, there is more room for capacity. When there's more room for capacity, your economy of scale goes up. So, if TCO improves, ROI improves, and your technical debt actually gets resolved a lot faster.

Gardner: It sounds as if managing the development, test and deployment cycle effectively really is almost like the head of a pyramid -- and affects the entire IT economic equation.

Naguib: Absolutely. There is a new paradigm now, they call it the "Escalator Message." In 60 seconds or less, we can talk to a CIO, CTO, COO, or CFO about our strategy and how we can help them shift from the firefighting mode to more of an architecture mode.

In 60 seconds or less, we can talk to a CIO, CTO, COO, or CFO about our strategy and how we can help them shift from the firefighting mode to more of an architecture mode.



If that’s the case, the more they can salvage their delivery, the more they can salvage their effective costs, and the more they can now shift to more of an IT-sensitive solutions shop. That helps build a business relationship and helps improve their economy of scale.

Gardner: We're hearing a lot here at Software Universe about ALM 11, a new launch by HP. You've been a beta user of at least some of the components of that. Tell us how that started and what you experienced?

Naguib: Sure. My background is that I dealt with the Mercury products back in the late 1990s. I have experience with Quality Center and the improvements that have gone on over the years. Because of our focus, we built our paradigm out of QA and into the performance world, and we started focusing on improving that process.

The latest TruClient product, which is a LoadRunner product, has been a massive groundbreaking point solution. In the last two years, frankly, with HP and Mercury getting adjusted, there’s been kind of a lag, but I have to give kudos to the team.

One of the key things is that they have opened up their doors in terms of the delivery, in terms of their roadmap. I've worked extensively for the last roughly year with their product development team, and they have done quite a bit of improvement in their solution.

Good partnership role

They have also improved their service support model; the help desk actually resolves questions a lot faster. And we also have a good partnership role, and we actually work with things that we see, and to the influence of their roadmap as well.

This TruClient product has been phenomenal. One of the key things we're seeing now is BPM solutions are more Ajax-based, and there are so many varieties of Ajax frameworks out there than we know how to deal with. One of the key things with the partnership is that we're able to target what we need, they are able to deliver, and we are able to execute.

Gardner: So, trying to fit that into our larger equation, the test and development deployment scenario is very important to the overall IT equation economically. How does this product, TruClient, fit into that in terms of aiding and abetting your goals?

Naguib: One of the key things is how to build partnerships across the organization, internally and externally. LoadRunner and TruClient allow us to get in front of the console, work with the business team, capture their typical use cases in a day-in-the-life scenario, and automate that. That gets buy-in and partnership with the business.

We're also able to execute a test case now and bring that in front of the IT side and show them the actual footprint from a business perspective and the impact and the benefits. What ends up happening is that now we're bringing the two teams together. So, we're bridging the gap basically from execution.

Frankly, nobody really cares as much about the footprint cost, until they start realizing the dollars that are spent.



Gardner: And as an early adopter and user, is there any 20-20 hindsight advice that you might offer to others who would be going down this trail as well?

Naguib: I would definitely send the message out to think in business value. Frankly, nobody really cares as much about the footprint cost, until they start realizing the dollars that are spent.

Also, now, business wants to see us more involved from the IT side, in terms of solutions, top-line improvements, and bottom-line improvements. As the performance teams expand and mature and we have the right toolsets, innovative toolsets like TruClient, we're able to now shift the cost of waste into a cost of improvements, and that’s been a huge factor in the last couple of years.

Last, I would say that in 8,000+ engagements -- we're actually closing in on now 10,000 events this year -- we've seen roughly $127 million in infrastructure savings that we have recouped. Again, that helps to benefit the firm. Instead of waste, now we're able to leverage that into more improvement side.

Gardner: So, the unfortunate reality for IT is they often have to do more with less. You've found a way to actually make that happen and perhaps continue it on an ongoing basis.

Naguib: Absolutely. I am excited about what I do. We have a great team and a great strategy. The support from my CEOs is fantastic. And again, we are seeing that just the whole partnership model across both the vendor side and internally has been a super benefit to the organization as well as the industry.

Gardner: Well, great. We've been discussing IT consolidation, applications lifecycle best practices with Abe Naguib, Director of Global Performance Architecture and Infrastructure Engineering at AIG Chartis insurance. Thanks so much, Abe.

Naguib: Thank you. I appreciate it.

Gardner: We're here in Barcelona at HP's Software Universe 2010 Conference. Look for this podcast and others on the HP.com website, as well as via the BriefingsDirect network.

I'm Dana Gardner, Principal Analyst at Interarbor Solutions, your host for this series of Software Universe Live discussions. Thanks for listening, and come back next time.

Listen to the podcast. Find it on iTunes/iPod and Podcast.com. Download the transcript. Sponsor: HP.

Transcript of a sponsored BriefingsDirect podcast, part of a series on application lifecycle management and HP ALM 11 from the HP Software Universe 2010 conference in Barcelona, Spain. Copyright Interarbor Solutions, LLC, 2005-2010. All rights reserved.

You may also be interested in: