Transcript of a BriefingsDirect podcast
with AIG and HP on the challenges and solutions involved in managing a
global center of excellence for IT performance.
Dana Gardner: Hello, and welcome to the next edition of the
HP Discover Performance Podcast Series. I'm
Dana Gardner, Principal Analyst at
Interarbor Solutions, your moderator for this ongoing discussion of
IT innovation and how it’s making an impact on people’s lives.
 |
| Dana Gardner |
Once again, we're focusing on how IT leaders are
improving performance of their services to deliver better experiences and payoffs for businesses and end-users alike.
We're now joined by our co-host and moderator, Chief Software Evangelist at HP,
Paul Muller. Welcome, Paul, how are you?
Paul Muller: I'm great Dana. How are you doing?
Gardner: I'm excellent. Where are you coming from today?
Muller:
I'm coming from sunny San Francisco. It’s unseasonably warm, and I'm really looking forward to today’s discussions. It should
be fun. [Disclosure:
HP is a sponsor of
BriefingsDirect podcasts.]
Gardner: We have a fascinating show. We're going to be learning about global insurance leader
American International Group, or AIG, and how their
Global Performance Architecture Group has leveraged a
performance center of excellence (COE) to help drive business transformation.
So let me introduce our guest from AIG. We're here with
Abe Naguib, Senior Director of AIG’s Global Performance Architecture Group. Welcome back, Abe.
Naguib: Hi, Dana. Hi, Paul. How are you?
Gardner:
We're excellent. We've talked before, Abe, and I'm really delighted to
have you back. I want to start at a high level. Many organizations are
now focusing more on the user experience and the business benefits and
less on pure technology, and for many, it's a challenge. From a very
high level, how do you perceive the best way to go about a cultural
shift, or an organizational shift, from a technology focus more towards
this end-user experience focus?
The CIO has to keep his eye forward to periodically change tracks,
ensuring that the customers are getting the best value for their money.
Naguib:
Well, Paul and Dana, there are several paradigms involved from the COO
and CFO’s push on innovation and efficiency. A lot of the tooling that
we use, a lot of the products we use help to fully diversify and resolve
some of the challenges we have. That’s to keep change running.
The
CIO has to keep his eye forward to periodically change tracks, ensuring
that the customers are getting the best value for their money. That’s a
tall order and, he has to predict benefit, gauge value, maintain
integrity, socialize, and evolve the strategy of business ideas on how
technology should run.
We have to manage quite a few
challenges from the demand of operating a global franchise. Our COE
looks at various levels of optimization and one key target is customer
service, and factors that drive the value chain.
That’s
aligning
DevOps to business, reducing data-center sprawl, validating
and making sense of vendors, products, and services, increasing the
return on investment (ROI) and
total cost of ownership (TCO) of emerging technologies,
economy of scale, improving services and
hybrid cloud
systems, as we isolate and identify the cascading impacts on systems.
These efforts help to derive value across the chain and eventually help
improve customer value.
Gardner: Paul Muller,
does this jibe with what you're seeing in the field? Do you see an
emphasis that’s more on this sort of process level, when it comes to IT
with of course more input from folks like the COO and the chief
financial officer?
Level of initiatives
Muller:
As I was listening to Abe's description I was thinking that you really
can tell the culture of an organization by the level of initiatives and
thinking that it has. In fact, you can't change one without changing the
other. What I've just described is a very high level of cultural
maturity.
 |
| Paul Muller |
We
do see it, but we see it in maybe 10 to 15 percent of organization that
have gone through the early stages of understanding the performance and
quality of applications, optimizing it for cost and performance, but
then moving through to the next stage, reevaluating the entire chain,
and looking to take a broader perspective with lots of user experience.
So it's not unique, but it's certainly used among the more mature in
terms of observational thinking.
Gardner: For
the benefit of our audience, Abe, tell us a little bit about AIG, its
breadth, and particularly the business requirements that your Global
Performance Architecture Group is tasked with meeting?
Naguib: Sure, Dana.
AIG
is a leading international insurance organization, across 130
countries. AIG’s companies serving commercial, institutional, individual
customers, through one of the world’s most extensive property/casualty
networks, are leading providers of life insurance and retirement
services in the US.
Among the brand pillars that we
focused on are integrity, innovation, and market agility across the
variety of products that we offer, as well as customer service.
Bringing together our business-critical and strategic drivers across
IT’s various segments fosters alignment, agility, and eventually unity.
Gardner: And how about the Global Performance Architecture Group? How do you fit into that?
Naguib:
With AIG’s mantra of "better, faster, cheaper," my organization’s
people, strategy, and comprehensive tools help us to bridge these gaps
that a global firm faces today. There are many technology objectives
across different organizations that we align, and we utilize various
HP solutions to drive our objectives, which is getting the various IT delivery pistons firing in the same direction and at the right time.
These include performance,
application lifecycle management (ALM), and
business service management (BSM), as well as
project and portfolio management (PPM).
Over time our Global Performance organization has evolved, and our
senior manager realized our strategic benefit and capability to reduce
cost, risk, and mitigate production and risk.
Our role eventually moved out of
quality assurance's QA’s functional testing area to focus on emphasizing
application performance,
architecture design patterns, emerging technologies, infrastructure and
consolidation strategies, and risk mitigation, as well increasing ROI
and economy of scale. With the right people, process, and tools, our
organization enabled IT transparency and application tuning, reduced infrastructure
consumption, and accelerated resolution of any system performances in
dev and production.
The key is bringing together our
business-critical and strategic drivers across IT’s various segments
fosters alignment, agility, and eventually unity. Now, our leaders seek
our guidance to help tune IT at some degree of financial performance to
unlock optimal business value.
Culture of IT
Gardner:
What's interesting to me, Paul, about what Abe just said is the
evolution of this from test and dev in QA to a broader set of first IT,
then operations, and then ultimately even through that culture of IT
generally. Is that a pattern you're seeing that the people in QA are in
the sense breaking out of just an application performance level and
moving more into what we could call IT performance level?
Muller:
As I was listening to Abe talk through that, there were a couple of
keywords that jumped out that are indicators of maturity. One of them is
the recognition that, rather than being a group-sized task, things like
application, quality performance, and user experience actually are a
discipline that can be leveraged consistently across multiple
organizational units and, whether you centralize it or make it uniform
across the organization is an important part of what you just described.
Maturity
of operational and strategic alignment is something that requires a
significant investment on business’s and IT’s behalf to prove early
returns by doing a good job on some of the smaller projects. This shows a
proven return on investment before the organization is typically going
to be willing to invest in creating a centralized and an uniform
architecture group.
Gardner: Abe, do you have some response to that?
Naguib:
Yes, more-and-more, in the last six or seven years, there's less focus
on just basic performance optimization. The focus is now on business
strategy impact on infrastructure
CAPEX, and
OPEX. Correlating business use cases to impact on infrastructure is the golden grail.
I always say that software drives the hardware.
Once
you start communicating to CIOs the impact of a system and the cost of
hosting, licensing, headcount, service sprawl, branding, and services
that depend on each other, we're more aligning
DevOps with business.
Muller:
You can compare the discussion that I just had with a conversation I
had not three weeks ago with a financial institution in another part of
the world. I asked who is responsible for your end-to-end business
process -- in this case I think it was mortgage origination -- and the
entire room looked at each other, laughed, and said "We don't know."
So
you've really got this massive gap in terms of not just IT process
maturity, but you also have business-process maturity, and it's very
challenging, in my experience, to have one without having the other.
Gardner:
I think we have to recognize too that most businesses now realize that
software is such an integral part of their business success. Being adept
at software, whether it's writing it, customizing it, implementation
and integration, or just overall lifecycle has become kind of the
lifeblood of business, not just an element of IT. Do you sense that,
Abe, that software is given more clout in your organization?
Naguib:
Absolutely Dana. I truly believe that. I've been kind of an internal
evangelist on this, but I always say that software drives the hardware.
Whether I communicate with the enterprise architects, the dev teams, the
infrastructure teams, software frankly does drive the hardware.
That's
really the key point here. If you start managing your root cost and
performance from a software perspective and then work your way out,
you’ve got the key to unlocking everything from efficiencies to
optimizing your ROI and to addressing TCO over time. It's all business
driven. Know your use cases. Know how it impacts your software, which
impacts your infrastructure.
Converged infrastructure
Gardner: Of course, these days we’re hearing more about software-defined networking,
software-defined data centers,
and converged infrastructure. It really does start to come together, so
that you can control, manage, and have a data-driven approach to IT,
and that fits into
ITIL
and some of the other methodologies. It really does seem to be kind of a
golden age for how IT can improve as performance, as productivity, and
of course as a key element to the overall business. Is that what you’re
finding too, Abe?
Naguib: Absolutely. It's targeting software performance, and
software-as-a-service (SaaS) applications that depend on each other.
More
and more, it's a domino effect. If you don't identify the root cause,
isolate it, and resolve it, the impact does have a cascading effect, on
optimization, delivery, and even cost, as we’ve seen repeatedly in the
last couple of years. That’s how we communicate to our
C-level community.
Gardner:
Of course we have to recognize it. Just being performant, optimized,
and productive for its own sake isn’t good enough in this economy. We
have to show real benefits, and you have to measure those benefits.
Maybe you have some way to translate how this actually does benefit your
customers. Any metrics of success you can share with us, Abe?
Naguib:
Yes, during our initial requirements-gathering phase with our business
leaders, we start defining appropriate test-modeling strategy, including
volumetrics, and managing and understanding the deployment pattern with
subscriber demographics and user roles. We start aligning DevOps
organizations with business targets which improves delivery
expectations, ROI, TCO, and capacity models.
The big transformation taking place right now is that our organization
is connecting different silos of IT delivery, in particular development,
quality, and operations.
Then, before
production, our Application Performance Engineering (APE) team
identifies weak spots to provide the production team with a reusable
script setting thresholds on exact hotspots in a system, so that
eventually in production, they can take appropriate productive measures.
Now, this is value add.
Gardner: Paul, do you
have any thoughts in terms of how that relates to the larger software
field, the larger enterprise performance field?
Muller:
As we’re seeing across the planet at the moment, there's a recognition
that to bring great software and information is really a function of
getting
Layers 1 through 7 in the technology stack working, but it's also about getting Layer 8 working.
Layer 8, in this case, is the people. Unfortunately, being technologists, we often forget about the people in this process.
What
Abe just described is a great representation of the importance of
getting not just a functional part of IT, in this case quality and
performance working well, but it's about recognizing the software will
one day be delivered to operational staff to internally monitor and
manage it in a production setting.
The big
transformation taking place right now is that our organization is
connecting different silos of IT delivery, in particular development,
quality, and operations, to help them accelerate the release of quality
applications, and to automate things like threshold setting, and
optimize monitoring of metrics ahead of time. Rather than discovering
that an application might fail to perform in a production setting, where
you've got users screaming at you, you get all of that work done ahead
of time.
Sharing and trust
You
create a culture of sharing and trust between development, quality, and
operations that frankly doesn’t exist in a lot of process where the
relationship between development and operations is pretty strained.
Gardner:
Abe, how do you measure this? We recognized the importance of the
metrics, but is there a new coin of the realm in terms of measurement?
How do you put this into a standardized format that you’re going to take
to your CFO and your COO and say here’s what's really happening?
Naguib:
That's a good question. Tying into what Paul was saying, nobody cared
about whether we improved performance by three seconds or two seconds.
You care at the front end, when you hear users grumbling. The bottom
line is how the application behaves, translating that into business
impact as well as IT impact.
Business impact is what
are the dollar values to make key use cases and transactions that don't
scale. Again, software drives the hardware. If an application consumes
more hardware, the hardware is cheap now-a-days, but licenses aren’t.
You have database and you have middleware products running in that
environment, whether it's on-premise or in the cloud.
The
point is that impact should be measured, and that's how we started
communicating results through our organization. That's when we started
seeing C-level officers tuning in and realizing the impact of
performance of both to the bottom line, even to the top line.
We were able to leverage consistent dashboards across different IT
solutions internally, then target weak spots and help drive optimization.
Gardner:
It strikes me, Abe, that this is going to set you up to be in a better
position to move to cloud models, consume more SaaS services, as you
mentioned earlier, and to become more of a hybrid services delivery shop
or have that capability. Does that make sense? Do you feel more
prepared for what this next level of compute architecture you seem to be
heading toward as a result of the investments you've made?
Naguib: Absolutely Dana. Our role is to provide more insight earlier and quicker to the right people at the right time.
Leveraging
HP’s partnership and solutions helped us to address technologies,
whether Web 2.0, client-server, legacy systems, Web, cloud-based, or
hybrid models. We were able to leverage consistent dashboards across
different IT solutions internally, then target weak spots and help drive
optimization, whether on premise or cloud.
Gardner:
Paul Muller, thoughts about how this is working more generally in the
market, how people who get a grasp on global performance architecture
issues like AIG are then in a better position to leverage and exploit
the newer and far more productive types of computing models?
Muller:
In the enterprise today, it's all about getting your ideas out of your
head and making them a reality. As Abe just described, most of the best
ideas today that are on their way into business processes you can
ultimately turn into software. So success is really all about having the
best applications and information possible.
Understand maturity
The
challenge is understanding how the technology, the business process and
the benefits come together and then orchestrating that the delivery of
that benefit to your organization. It's not something that can be done
without a deliberate focus on process. Again, the challenge is always
understanding your organization's maturity, not just from an IT
standpoint, but importantly from a broader standpoint.
Naguib:
What's the common driver for all? Money talks. Translating things into a
dollar value started to bring groups together to understand what we can
do better to improve our process.
Gardner: Abe,
it strikes me that you guys are really fulfilling this value epicenter
role there and expanding the value of that role outside the four walls
of IT into the larger organization. Tell me how HP is joining you in a
partnership to do that? What is it that you're bringing to the table to
improve that value for the epicenter of value benefit?
Naguib:
Dana, what we're seeing more is that it's not just internal dev and ops
that we're aligning with, or even our business service level
expectations. It's also partnerships with key vendors that have opened
up the roadmap to align our technologies, requirements, and our
challenges into those solutions.
The gains we make are
simple. They can be boiled down into three key benefits: savings,
performance, and business agility. Leveraging
HP's ALM
solutions helps us drive IT and business transformation and unlock
resources and efficiencies. That helps streamline delivery and an
increased reliability of our mission critical systems.
After we've dealt with tuning, we can help activate post-production
monitoring using the same script, understanding where the weak spots
are.
My favorite has always been
HP's LoadRunner Performance Center.
It’s basically our Swiss Army Knife to support diverse platform
technologies and align business use cases to the impact on IT and
infrastructure via SiteScope, HP SiteScope.
We're able
to deep dive into the diagnostics, if needed. And the best part is,
after we've dealt with tuning, we can help activate post-production
monitoring using the same script, understanding where the weak spots
are.
So the tools are there. The best part is integrated, and actually work together very well.
Gardner: It really sounds like you've grabbed onto this system-of-record concept for IT, almost
enterprise resource planning (ERP) for IT. Is that fair?
Naguib: That's a good way to put it.
Muller:
One of the questions I get a lot from organizations is how we measure
and reflect the benefit. What hard data have you managed to get?
Three-month study
Naguib: IDC
came in and did an extensive three-month study, and it was interesting
what they have found. We've realized a saving of more than $11 million
annually for the past five years by increasing our economy of scale.
Scale on a system allows more applications on the same host.
It's
an efficiency from both hardware and software. They also found that our
using solutions from HP increased staff productivity by over $300,000 a
year. Instead of fighting fires, we're actually now focusing on
innovation, and improving business reliability by over $600,000 a year.
So
all that together shows a recoup, a five-year ROI, about 577 percent. I
was very excited about that study. They also showed that we resolved
mean time resolution over 70 percent through production debugging, root
cause, and resolution efforts.
So what we found, and
technologists would agree with me, is that today, with hardware being
cheaper than software, there is a hidden cost associated with hosting an
application. The bottom line, if we don’t test and tune our
applications holistically, either the architecture, code,
infrastructure, and shared services, these performance issues can
quickly degrade quality of service, uptime, and eventually IT value.
I have a saying, which is that quality costs money but bad quality costs more.
Muller: I have a saying, which is that quality costs money but bad quality costs more. There you go.
Gardner:
Abe, any recommendations that you might have for other organizations
that are thinking of moving in this direction and that want to get more
mature, as Paul would say. What are some good things to keep in mind as
you start down this path?
Naguib: Besides
software drives the hardware -- and I can't stress that enough -- are
all the ways to understand business impact and translate whatever you're
testing into the business model.
What happens to the
scenarios such as outages? What happens when things are delayed? What is
the impact on business operability, productivity, liability, customer
branding. There are so many details that stem from performance. We used
to be dealing with the "Google factor" of two-second response time, but
now, we're getting more like millisecond response, because there are so
many interdependencies between our systems and services.
Another
fact is that a lot of products come into our doors on a daily basis.
Modern technologies come in with a lot of promises and a lot of
commitments.
Identify what works
So
it's being able to weed through the chaff, identify what works, how the
interdependencies work, and then, being able to partner with vendors of
those solutions and services. Having tools that add transparency into
their products and align with our environment helps bring things
together more. Treating IT like a business by translating the impact
into dollar value, helps to get lined up and responsive.
Gardner:
Very good. Last word to you, Paul. Any thoughts about getting started?
Are there principles that you are seeing in common, threads or themes
for organizations, as they begin to get the maturity model in place and
extend quality and process performance assurance improvements even more
generally into their business?
Muller: It might
be a little controversial here, but the first step is look in the mirror
and understand your organization and its level of maturity. You really
need to assess that very self-critically before you start. Otherwise,
you're going to burn a lot of capital, a lot of time, and a lot of
credibility trying to make a change to an organization from state A to
state B. If you don’t understand the level of maturity of your present
state before you start working on the desired state, you can waste a lot
of time and money. It's best to look in the mirror.
The
second step is to make sure that, before you even begin that process,
you create that alignment and that desired state in the construct of the
business. Make sure that your maturity aligns to the business's
maturity and their goal. I just described the ability to measure the
business impact in terms of revenue of IT services. Many companies can’t
even do something as fundamental as that. It can be really hard to
drive alignment, unless you’ve got business-IT alignment ahead of time.
I
have said this so many times. The technology is a manageable problem,
Layers 1 through 7, including management software to a certain degree,
have solved problems the most time. Solving the problem of Layer 8 is
tough. You can reboot the server, but you can’t reboot a person.
Solving the problem of Layer 8 is tough. You can reboot the server, but you can’t reboot a person.
I
always recommend bringing along some sort of management of
organizational change function. In our case, we actually have a number
of trained organizational psychologists working for us who understand
what it takes to get several hundred, sometimes several thousand, people
to change the way they behave, and that’s really important. You’ve got
to bring the people along with it.
Gardner: Well we
have to take a hint from you, Paul. Maybe our next topic will be The
Psychology of IT, but we won’t be able to get to that today. I am afraid
we'll have to leave it there and I have to thank our co-host Paul
Muller, the Chief Software Evangelist at HP. Thanks so much for joining
us.
Muller: Always a pleasure.
Gardner: And like to thank our supporter for this series, HP Software,
and remind our audience to carry on the dialogue with Paul and other
experts there at HP through the Discover Performance Group on LinkedIn.
You
can gain more insights and information on the best of IT performance
management at
www.hp.com/go/discoverperformance. And you can always
access this in other episodes of our HP Discover Performance podcast
series at hp.com and on iTunes under BriefingsDirect.
Of
course, we also extend a big thank you to our guest. Abe Naguib, Senior Director of AIG’s Global Performance Architecture Group.
Thanks so much, Abe.
Naguib: Thank you, Dana, thank you, Paul. I really appreciate the opportunity.
Gardner:
Again, a last thank you to our audience for joining us for this special
HP Discover Performance podcast discussion. I'm Dana Gardner, Principal
Analyst at Interarbor Solutions, your co-host for this ongoing series
of HP-sponsored business success story. Thanks again for joining and
come back next time.
Transcript of a BriefingsDirect podcast
with AIG and HP on the challenges and solutions involved in managing a
global center of excellence for IT performance. Copyright Interarbor
Solutions, LLC, 2005-2012. All rights reserved.
You may also be interested in:
We chose HP’s Real User Monitoring (RUM) and HP’s Operations Management i (OMi)
to help us decrease our mean time to repair and to obtain a better
understanding of end-user performance and how our applications are used
by our customers.
