Thursday, January 5, 2012

Travel giant TUI Group leverages virtualization management tools to drastically improve IT performance troubleshooting

Listen to the podcast. Find it on iTunes/iPod. Read a full transcript or download a copy. Sponsor: VMware.

Better managing virtualized IT workloads and private clouds is a top concern for IT leaders going into 2012. They may want to follow the lead of global travel and tourism giant TUI Group. The IT organization there, TUI InfoTec, has found ways to manage highly virtualized IT operations better, especially in mixed environments like hybrid clouds.

The critical need to better identify performance issues and outages prompted TUI InfoTec to find ways to cut time to troubleshooting, resulting in a 50 percent reduction in the time needed to identify the causes of such problems.

To learn more about better systems management in heterogeneous cloud environments and in virtualized environments, BriefingsDirect interviewed Christian Rudolph, Infrastructure Architect at TUI InfoTec in Hanover, Germany. The discussion is moderated by Dana Gardner, Principal Analyst at Interarbor Solutions. [Disclosure: VMware is a sponsor of BriefingsDirect podcasts.]

Here are some excerpts:
Rudolph: We're a very silo-based environment. We have dedicated network storage and a server team responsible for resolving issues in our infrastructure.

What we've seen in the past were a lot of problems in getting these people together. Everybody had different management tools from the different vendors and nobody had an over-all view about the infrastructure.

We're also starting to take a look at how, from a cost perspective, we can do the best for our customers.

We’re 60 percent in the Windows environment, and 20 percent in the UNIX environment, which is virtualized, and we're currently planning to go further -- to 80 percent virtualization in the total landscape. That's our current state, and we’ve driven more and more to a virtualized infrastructure for all the mission-critical systems.

Normally when we have performance issues, our responsibilities are not very clear -- this is a server problem, a network problem, an OS system problem, or this is only the end-user who has a problem. He feels that the application isn't fast enough. In the past, we had a large problem getting information all together.

This is where we evaluated VMware vCenter Operations to get an over-all overview about our infrastructure and to get a deep dive into our infrastructure to take a look at how can we solve problems faster and how this could help us in the normal process.

Now we have vCenter Operations on a single pane of glass that can roll down to the storage network and also the infrastructure CPU memory resources to have a clear overview of what could be the first root cause of an issue or performance for the end user. We've tried to figure out how can we bring it better together, and for us vCenter Operations, it’s a single pane of glass.

We currently use the vCenter Operations 1.0 Standard version, but we're in the beta program currently for 5.0. It's a new version, which comes out [in 2012] with vCenter Operations 5.0. These version give us the ability to do capacity planning and also performance analysis in one view so that we can adapt the things we have discovered in normal business hours for the system and also to do capacity planning for the future.

Gardner: Tell me a little bit about TUI, and TUI InfoTec.

Rudolph: TUI InfoTec is an external IT provider for the TUI AG Group. The TUI AG Group is a European leading company in travel and tourism. They're very large in Germany, in the UK, and also in other European countries. They’re not presently doing a lot of business in the US.

We started as an internal IT organization from TUI Germany, and moved in 2006 to an external service provider for the TUI AG and other companies. We're a joint venture company with Sonata Software Ltd., which holds about 50 percent of the company. We're responsible for all the business-critical IT for TUI AG group like the booking systems, the access planning system, and all the other systems related to the business of the TUI AG group.

If it comes to an outage of the IT systems we lose a lot of money. So we have to take care that everything is working and running in the infrastructure.

Gardner: How is your landscape for cloud?

Rudolph: We’re currently thinking about planning our private cloud for our development team. We're also starting to take a look at how, from a cost perspective, we can do the best for our customers. Maybe we can include peak trading for some of the systems. We have a great opening for producing catalogs for the customer, so that they're able to connect our internal cloud over to external clouds and have the hybrid clouds then in place.

Gardner: Okay. How has that beta with vCenter Operations 5.0 worked out? Are some of these features something that you think will be of value to you?

A good overview

Rudolph: We have two or three good cases there. This has really helped us in the normal business. We've been running with the beta for two months and what we've detected is that we have a good overview, because we have some multi-vCenter environments. We have, in total, three productive vCenters and we need to discover all of them. We had a problem, because we can't use Linked Mode for the vCenters. We had no central view for all the systems to get a performance overview of the system.

And there is a second step. We didn't have the capacity in the same view. So we weren't able to do capacity planning, until we manually got all the information from the different vCenters to have a consolidated planning view. For us, this is one of the most important things that we can do for planning in one place for all our vCenters and also know how many capacity hours are left for new machines. So we increased our time to deliver a virtual machine (VM).

Gardner: What has this better IT visibility in operations and remediation brought to you in technical and in business terms?

Rudolph: The process is very easy, because we've seen that we reduced the time until we can deliver our root cause for our known problem by nearly 50 percent. We reduced the time for doing that, and this is also the best case for our customers -- that we can deliver faster solution for a system problem.

The second thing we've seen is that we can see earlier information about how the system is feeling. Through vCenter Operations and through the health status in the vC Ops we can see how our end-users feel. We can detect some problems before they occur, and that’s the best use case we can ever have.

When we detect problems faster and can resolve them faster, they have faster usage of the product.

Gardner: How about looking toward the future? We talked a little bit about your use of improved operations, but will this become important when you move to more cloud, software-as-a-service (SaaS), and/or mobile types of activities. How important is this proactive ability in management as you innovate?

Rudolph: It's very important for us. We currently have the vCenter orchestration platform implemented, and we're starting to deliver to the end-user a service portal. Where they can request more-and-more VMs. When we didn’t have the products to monitor this system and we come to great trouble. How can we else go further, maybe to a hybrid cloud environment, if we can’t manage our private cloud like now with the vCenter Orchestrator and also with the vC Ops.

Gardner: Taking a step back and reviewing how things have gone, do you have any recommendations or advice for other companies that might be pursuing higher levels of virtualization and perhaps looking for similar reduction in meantime to solution for problems?

Two recommendations

Rudolph: I see two recommendations. Not many people know how powerful vCenter Orchestration is. This is one powerful tool as an automatic way for deployment, for maintaining, and also to do some other basic tasks in your virtual infrastructure. This is one important step for us to go to a higher virtualization ratio, because it can be delivered faster to our end-users.

The second thing is really to take a look at vCenter Operations and definitely to the new version that’s coming up. This really helps us to understand how my infrastructure is working. When I don’t know that, I may have problem with one of my disks and I/O and this reflects back to one VM especially. You have to know that, otherwise you don’t have recognition from the end-user that virtualization is really working and that you can bring mission-critical systems to the virtual infrastructure.
Listen to the podcast. Find it on iTunes/iPod. Read a full transcript or download a copy. Sponsor: VMware.

You may also be interested in:

No comments:

Post a Comment