Every company has its enterprise network management software ninjas who launch into action when alerts go off or...
users complain – not to ensure an internal service level agreement but to determine whether the problem is a network issue at all.
In fact, every IT silo -- servers, applications, storage -- has these ninjas running their own tools in isolation from one another, all hoping to prove the problem is someone else's fault.
In the meantime, crucial applications perform poorly. So, for many enterprises, this finger-pointing just isn't good enough anymore. That's why network management software vendors and IT management vendors are moving toward service assurance strategies in which all IT management tools, including enterprise network management software, feed into a single service delivery management console that gives applications teams, IT executives and business unit managers a unified view of how the entire IT infrastructure works together to deliver an application. When something goes wrong, the IT silos can work together to resolve the issue rather than retreating to their individual toolsets.
CA Technologies has reorganized much of its IT management portfolio into a service delivery management product line in which many of its tools, including recently acquired NetQoS Performance Center, roll into a CA Spectrum Service Assurance console. Today, NetScout introduced nGenius Service Delivery Manager, which takes the network traffic collection capabilities of the company's enterprise network management software and gives customers a high-level view of how different components of IT infrastructure are working together to deliver applications. NetScout has also designed the product to integrate with some of the major IT management vendors' products, including IBM Tivoli and HP Openview.
How one service delivery management model works
For the last two years, Craig Yappert has been leading the effort to institute a service delivery management model at Oracle, where he serves as senior director of enterprise monitoring solutions for the company's global IT organization. He has been working to combine network management tools from Monolith Software and NetScout with Oracle's own Enterprise Management product and an internal configuration management database (CMDB) into a custom-built service assurance platform. The project has led to a transformation of how Oracle troubleshoots its internal applications.
"In the past … a help desk ticket … would go to the team responsible for the application, and they would begin troubleshooting. The server team would go off and try to solve the problem. The application team would try to solve the problem. The storage team would try to solve the problem and the network team would try to solve the problem," Yappert said. "Having a consolidated view of what is going on internally reduces the time and the complexity of communication [between silos]…. They're working from an application layer into a network layer, and they can say, 'The reason we're having a problem is because the firewall is overloaded.'"
Service delivery management also enables network managers to cull trends from network monitoring data in a more proactive manner.
"We can say from a capacity perspective, this is what your service looks like," Yappert said. "We're seeing a lot of network activity increase from month to month, and we're seeing an increase in storage activity. We see that you're moving a lot of data from storage to your application layer across the network. We need to optimize that. Maybe we need to look at a new architecture, or we can look at partitioning you off and giving you more bandwidth."
Network managers should be ready for a service assurance initiative or internal SLA
Networking teams have to keep their eye on the emergence of service assurance models in their companies, since they typically don't drive the shift toward service delivery management.
"The network guys are typically the last ones invited to the party," said Jim Frey, research director for Enterprise Management Associates. "Service assurance usually comes from the CIO down. It usually starts at the help desk and moves into application support and then into the data center. Then the network guys get brought in after that."
Network managers shouldn't sit and wait for it to happen, however. They should make sure that the enterprise network management software they have today will fit inside a service assurance model.
"Every network monitoring tool over time needs to understand how to present the data they collect in a business-oriented or service-oriented manner," Frey said. "You've got to be able to provide information back to the operations team that says what services are affected [by an event] and what business group this is affecting. Those that are part of a broader suite can do it intrinsically. Those that aren't have to integrate with someone who has a service assurance model."
Yappert recently replaced a legacy network monitoring product with a product from Monolith Software because it had open APIs that Oracle could use to integrate the product into its overall service delivery management technology.
"When we select vendors now, and what we are driving existing vendors to do, is to ensure they provide us with open APIs so we can access information and repurpose that information to fit the specific needs of the business rather than changing our business needs to fit the tool," he said.
Service delivery management begins with the right tools
Jason Hall, director of IT systems for furniture retailer Rooms To Go, is a CA customer who uses many of the enterprise network management software and application management software products that make up CA's service assurance package. His company uses CA NSM, CA Spectrum Infrastructure Manager and CA eHealth Performance Manager. Hall is not quite ready to add the Spectrum Service Assurance product to tie it all together because he simply has too many other projects on his plate at the moment.
However, he recently added NetQoS Performance Center to fill a gap in the company's ability to manage and monitor application performance and bandwidth utilization, which has given him the basic tools to start managing infrastructure from a service delivery management perspective. Before installing NetQoS, Hall had no visibility into how end users were experiencing application and service performance across the WAN or LAN.
"It was purely the end user," he said. "We waited for someone to call. Operationally, that gives the end user the perspective that the systems are slow … and that we're not doing anything about it. It decreases their confidence in the actual systems. Sneakernet is never a reliable source of information. Once it gets to the third person, the information is not that good anymore. You end up chasing phantoms real quick."
Hall said that adding NetQoS's performance management capabilities to his suite of tools has helped him monitor the delivery of critical centralized applications, such as a point of sale application that is used by all of his company's stores. It's also helped him solve some service delivery mysteries, particularly with his company's intranet.
"We discovered that the server [hosting the intranet] was pretty much awful," he said. "We'd had people complain. The developers looked at the application and said, 'We don't see anything.' We went back and forth with them."
Hall's team used NetQoS to demonstrate that the network was performing just fine when delivering intranet applications, and they were able isolate the problem to the server.
Ultimately, Hall wants to combine all his management tools into a single service assurance console. That way, he will avoid jumping from console to console to find the answers.
"That's exactly it … being able to get the answer quicker with a unified view and being able to tie it all together instead of having the pieces and the slices spread around," he said. "That is a place we need to head to, but right now I just don't have the time to undertake another project."
Let us know what you think about the story; email: Shamus McGillicuddy, News Editor