Delivering End-to-End Visibility Across the Hybrid Enterprise

Searching for the source of application Performance problems

IT complexity increases when companies use cloud-based and SaaS apps together with on-premises apps. And complexity increases further when companies use public Internet bandwidth along with private MPLS networks to deliver those apps. All of this together creates a hybrid state of IT beset with many blind spots that block visibility into quality of service and performance issues. As a result, IT engineers spend 10-50% of their time on root-cause analysis, according to Forrester Research, endlessly searching for the sources of application performance problems in a sea of complexity.

Another critical issue with lack of visibility arises when companies develop apps in the cloud. To develop applications faster and more efficiently, development teams are using large numbers of third-party components as building blocks.

The average enterprise downloads more than 229,000 such third-party components annually, according to Sonatype, which runs the central repository for many Java software building tools. Such components account for 80% to 90% of the code in a typical enterprise application today. They also add a lot of opaqueness to the internal operations of the application.

Nearly two-thirds (64%) of enterprises continue to use a fragmented approach to technology monitoring, which cannot possibly deliver end-to-end visibility. ¹ The typical enterprise has 6 – 10 network monitoring and troubleshooting tools in use, and 10% of large enterprises have more than 25. ²

But, because each domain has a different perspective and restricted view, engineers often arrive at conflicting conclusions in war rooms as they’re trying to solve application performance problems that are crippling the business. The fragmented approach to performance problems and lack of unified, cross-domain visibility leaves IT stuck in siloes, reactive rather than proactive, wasting time finger-pointing, with performance issues too often unresolved and continuing to impact the business.

For example, SAS, the leader in business analytics software and services, supports equipment in more than 100 locations around the world. Many of the company’s business-critical applications, including CRM, VoIP, and training systems, are centralized and delivered to offices over a hybrid WAN. In addition, SAS employees are using more and more SaaS applications.

“As SAS has grown we have globalized our IT operations,” says Doug Bradley, senior information systems engineer at SAS. “While that means greater complexity, it also opens the opportunity to centralize and streamline application performance management. In the past, when application problems occurred, the IT team would try to recreate the issue and observe what was happening in real time. We wanted a solution that would speed results and support more efficient use of our resources.”

Riverbed solutions now give Bradley’s team end-to-end visibility into application and network performance. “With Riverbed, issues that might have taken weeks previously take only days now. With the data stored in the appliances, we can just dial back in time. It’s like a time machine,” says Bradley. “We appreciate the ability to see key performance indicators at a glance.”

Southeastern Grocers can NOW see what’s going on with their cloud apps

Another critical issue with lack of visibility arises when companies develop apps in the cloud. To develop applications faster and more efficiently, development teams are using large numbers of third-party components as building blocks.

The average enterprise downloads more than 229,000 such third-party components annually, according to Sonatype, which runs the central repository for many Java software building tools. Such components account for 80% to 90% of the code in a typical enterprise application today. They also add a lot of opaqueness to the internal operations of the application.

For example Southeastern Grocers discovered that when developing a new app to deliver personalized digital coupons and other perks directly to customers’ smartphones, with the goal of increasing customer loyalty by personalizing the shopping experience. “We lost a lot of visibility into our infrastructure when we went to the cloud and that was kind of challenging at first,” says Darren McDaniel, IT lead for marketing at Southeastern Grocers.

Although most of Southeastern Grocers’ data and applications are housed in its data center, the new mobile app was built in the cloud using Microsoft Azure, a coupon platform from Coupons.com, and third-party services in addition to home-grown code. “We’ve used some third-party services and some internal development. It’s a hodgepodge of everything that we’re trying to make into one unified experience for the customer,” says McDaniel. “We used a middle layer gateway in the SOA, and we had found that this gateway wasn’t disposing old connections if there was an error. So we were running out of connections, and the connection pool would die and stop accepting connections.”

“It’s a small piece of code,” says McDaniel, “but the application framework is running under it. And the framework was making 1,500 calls to the database during these actions when it shouldn’t, meaning that the thread was still open for another 15 seconds after that user finished. We were discouraged, because we weren’t able to break it down into the actual line of code inside of the system so the vendor could issue a fix.”

McDaniel fixed the issue after working with Teneo to monitor applications deployed on Microsoft Azure Cloud Services and enable visualization of application behavior in real-time. With Riverbed-enabled x-ray visibility into every aspect of performance, “We were able to prove the issue and take it back to the gateway vendor team,” says McDaniel. “We very seldom have issues anymore. Riverbed has helped us to narrow those down and fix the issues before they become an issue.”

That’s true visibility: The ability to see and understand what’s going on with your apps and services with clarity across the entire enterprise from cloud to datacenter to branch to user, giving you the insights you need to enforce SLAs across SaaS apps, cloud-based apps, and on-premises apps and infrastructure, and to make the best business decisions with confidence that IT can support them.

For more information we recommend reading “The 4 Stages of Cloud Adoption”

[1] Forrester Research, “Application Performance Management Is Critical To Business Success,” February 2014.

2 EMA, “Managing Networks in the Age of Cloud, SDN, and Big Data: Network Management Megatrends 2014,” April 2014.