Network Performance

Rapid response benefits county

Application performance issues are addressed by a problem identification appliance.

Paul Roybal stepped into his role as the CIO of Bernalillo County, N.M., just over two years ago, at a time when the county's IT department did not have an exemplary reputation among employees. Roybal's mission was to move the organization from being a technology–driven unit into one that was delivering direct results for the end–users it supported. To do so, the IT team needed to focus on the areas of innovation that would have direct business impact on productivity and customer satisfaction.

During this same time, Roybal was tasked with moving the "internal" IT shop to one that also delivers services that can be leveraged by every member of the community, which includes Albuquerque. "Information technology in the public services sector has to tackle the growing challenge of satisfying twin constituencies," notes Roybal. "In our case, that means our employee customer base, as well as, potentially, 600,000–plus county residents. As a result, my organization is continuously focused on developing best practices and acquiring best–of–breed solutions that enable us to extend our service orientation to a rapidly expanding end–user community."

The team had to implement its innovations creatively within a fixed budget, but Roybal contends the county has managed to find a balance between leading–edge technology and return on investment. "We've made substantial investments in server and storage virtualization, VoIP, data warehousing, and the like," he says, "but we've always placed the primary emphasis on how these solutions will make county employees more productive and responsive to the members of our community. On these fronts, we've succeeded to a great degree.

With RPI, the team was able to get a holistic view of the entire Bernalillo environment with auto–discovery of the endpoints, applications and networks across silos.

"But another secret to our success is to ensure that the complexity of what we've done is hidden from our end–users. To achieve that goal, we have to have management solutions in place that keep pace with our rate of innovation."

Initially, Roybal notes, his team put together an applications–delivery architecture, moving all cross–departmental applications into a centralized client–server model. This migration improved cost efficiencies and user productivity, but IT also started to see a number of cases of intermittent degradation of application performance. Finding a solution that could discover, manage and resolve network problems that cause degraded application performance was critical because the problem could potentially be compounded by the county's planned migration to a new enterprise resource planning (ERP) solution from SAP.

SECURITY PRODUCT NOT ENOUGH

Roybal and his team explored a few different options and then settled on a security product based on flow technology, hoping it would help the network operations team get to the heart of the performance–degradation problem and other complex issues much more quickly. These flow–based tools were optimized to capture virus outbreaks so Roybal's team would be notified when there was an outbreak. There was no way to determine, however, if there was an issue between an application and the network and how that impacted performance.

"The security products allowed us to monitor user interactions with the network, but these tools were missing an understanding of how the networks, applications and endpoints interact with each other," says Roybal. "Without that focus, we couldn't pinpoint the source of the performance problem and tackle productivity issues in the way that we had hoped."

Roybal and his team once again began their hunt for a solution that could resolve these performance issues. The team found that some of the more traditional solutions typically supply IT with a mountain of data that can take hours to analyze–and, even then, IT may not have the proper evidence to solve problems.

At this critical juncture, Bernalillo County was introduced to the Xangati rapid problem identification (RPI) appliance and put it through an evaluation process designed to assess its value relative to the application performance issues. The Xangati appliance is designed to shrink the time between the genesis of a problem that can affect application performance and the identification of the problem's source.

With RPI, the team was able to get a holistic view of the entire Bernalillo environment with auto–discovery of the endpoints, applications and networks across silos. This gave the team an integrated understanding of all the elements that make up the application from the network itself to all the various endpoints. The RPI system gave Roybel's team a "navigable window" into all the interactions that were taking place in their environment, enabling them to zoom in and zoom out of a problem.

During the initial evaluation of the product, its precision endpoint and application profiling model was able to direct the IT team to the sources of the intermittent performance problems–a hyperactive endpoint that was driving down mission–critical application performance for other endpoints that had to access the same central server resources.

PROBLEM SOURCE IDENTIFIED

"We had other products that let us monitor interactions, but they didn't allow us to get out in front of the problem," explains Roybal. "My team and I were pleased to see immediate results from the solution that helped us identify in real time the source of a complex application responsiveness issue."

Because the RPI model leverages flow data from the county's installed base of routers, it is interoperable with other network components and does not disrupt network configuration or traffic running across the network. The appliance installed quickly, without requiring costly software, agents or hardware probes. The implementation team simply set a CLI command on the routers to generate flows and installed the system in less than 30 minutes via its user interface.

Because IT saw that the Xangati appliance provided a marked difference in the actionable information it provided, the county moved quickly to deploy the RPI solution. On the first day, IT found that the bandwidth–consumption characteristics of a backup application process were resulting in intermittently sluggish performance with other centralized applications that one department leveraged over a WAN connection.

In another case, the RPI solution found that loads for a number of the county's key applications were not distributed evenly across the key server pool. IT was able to improve the performance of these applications by tweaking configurations on the load balancers. Among the other problems the solution helped uncover were a sluggish e–mail server, an intermittent domain name system failure and an endpoint hijacked for spamming.

"In our initial cases, we were already seeing our problem identification cycles shrink by at least 33 percent and staff productivity has increased," says Roybal. "Our analysis showed a payback on the product less than six months after deployment."

Now Bernalillo County is using the RPI solution within centralized help desk support processes, where it helps enhance the IT team's responsiveness to customer support calls. IT can now better qualify user complaints and resolve issues more readily because personnel get real–time views into what end–users are doing with their networked applications. Now problems are assigned more quickly to the right subject–matter expert, and Xangati's RPI correlation capabilities allow IT personnel to close trouble tickets for complex issues 20 percent faster than before.

For more information from Xangati (click here)