Chasing the Network Visibility Problem
IT admins can lose control of processes if the virtualized parts of their networks aren't monitored correctly. Here are the issues and some solutions.
- By Dimitri Vlachos
The corporate network connects employees, customers, business partners, and supporting companies in day-to-day operations, and serves strategically as the platform for business growth. This central IT asset is evolving alongside trends, such as cloud computing, virtualization, and mobile computing. As they adopt these new technologies, organizations struggle to balance flexibile access with control over data and costs. They'll often retrench and move data, storage, and servers to centralized environments, and/or turn to virtualization to lower costs, consolidate resources, and increase efficiency. Regardless, this means files and applications are moving farther away from users, resulting in network visibility problems at a time when users expect increased levels of reliability, stability, and performance from the network, whether they are in the office, on the road, or working from home.
Virtualization: New Network Visibility Problems
Increased reliance on network connectivity spells more calls to the IT help desk, as any lapse in data accessibility can frustrate users. The network is usually the first to be blamed, often unfairly. And now as more businesses virtualize their critical applications, they create additional visibility gaps in the IT infrastructure, and it becomes even more difficult to pinpoint where issues are occuring and what is causing them.
Traditional monitoring tools can only monitor the physical network. Once an application goes into the virtualized server, the IT organization will lose track of what is happening. To remedy this, IT organizations need a virtual tap that monitors traffic flowing over the virtual switch and forwards the information to a centralized analysis solution that presents end-to-end visibility into the application -- across both the virtual and physical network -- from the same dashboard.
While the network is sometimes the culprit, a slowdown in application or network performance can often be caused by an unrelated issue, resulting in more frustration and tail-chasing as network managers look to identify the cause of a problem that may not actually exist.
IT organizations would love to have complete visibility into their networks and applications so that they can quickly identify where the problem is occurring and resolve them, preferably before they impact the business. Unfortunately, with traditional network management solutions, IT may only become aware of an issue once a user calls the help desk, and by the time they identify its source and set about fixing it, user frustration is often running high.
The visibility problem is exacerbated for many organizations that have adopted virtualization or private cloud deployments. While virtualization can bring many benefits to the organization with improved resource utilization, higher availability, and capital savings, it also brings a management complexity that threatens to diminish the benefits, especially with added operational costs.
It is no wonder that blame is targeted at the network when problems arise; it is easy to do. With so many new demands and more end points to the network, IT managers may have little idea of what is actually running on their networks -- specifically, the applications, servers, and the employees who use the networks, as well as the dependencies that exist between them. This creates difficulties in monitoring and troubleshooting, particularly when planning IT projects such as consolidation, disaster recovery, or cloud-based computing. It is often simply impossible to make network changes without disruptions, or to diagnose network problems unless IT assets are properly discovered and mapped.
The Solution: Application-Aware Network Performance Management
There is a better way for IT organizations to stay on top of their networks, and it calls for proactive network management based on network behavioral analytics and application performance monitoring. When combined with an efficient discovery and dependency mapping process, organizations can quickly identify and respond to IT performance issues before they impact the business.
Application-aware network performance management (NPM) solutions use analytics capabilities to automatically baseline and track dozens of performance metrics, and when meaningful changes are made, send proactive alerts to IT, which can often respond before users become aware of degradations in application or network performance. In this environment, when concerns are reported, the source of the issue can be quickly identified, allowing IT managers to focus their efforts to resolve it.
Flow-based discovery and dependency mapping programs are another important piece of successful network performance management. Discovery is used to map IT assets and their dependencies. It identifies all the components -- application servers, web servers, load balancers, databases, authentication servers, network components, etc. -- used to deliver an application as a service to the end user.
Understanding application dependencies not only enhances the ability to locate the source of performance bottlenecks, but also means that IT consolidation, virtualization, cloud initiatives, and other changes to the network can be made with confidence.
ROI from Improved Visibility: Quantifying Results
Analyst firm IDC recently conducted a study which identified that after a company implemented network performance management tools, on average it achieved benefits in three areas: IT infrastructure cost reduction ($12,047 per 100 users from optimizing bandwidth and network resources, server consolidation, and increased automated network management), IT staff productivity improvement (the equivalent of 5.4 full-time employees as a result of identifying and resolving networking and application problems more efficiently -- a savings of $5,497 annually per 100 users), and increased end-user productivity (a savings of $26,057 annually per 100 users). Also, the average time taken to resolve an IT incident was typically reduced by 67 percent, while help desk call duration was cut by 92 percent.
By automating dozens of key performance metrics on a per-site and per-application basis, IT managers gain an early warning system of unexpected changes in behavior. Now, when issues arise, the relationship between applications and infrastructure is readily understood, accelerating the problem identification and its resolution. And, whenever the network is blamed for problems, IT managers are empowered to redirect it to the actual source. Furthermore, a comprehensive dashboard can provide more than just clear monitoring; it can also help with long-term planning and establish fair use network policies, which can significantly benefit end users and their organizations alike.
Commit to Seeing the Entire Network Picture
Corporate networks are evolving along with industry trends, such as cloud computing, virtualization, bandwidth-intensive or latency-sensitive applications (e.g., voice and video), and mobile computing. Managing a corporate IT network and maintaining high performance and security is an increasingly challenging task but it should not be a tail-chasing one.
Application-aware NPM solutions combine network and application monitoring to help IT organizations address problems almost instantly, better allocate network resources, and help ensure IT is closely aligned with business and operational goals. Now, with an end-to-end view into application performance as it traverses both the physical and virtual network, network managers can do less tail-chasing and more tail-wagging.
Dimitri Vlachos is vice president of marketing and products, Cascade Business Unit at Riverbed Technology.