The network fallout every CIO should watch out for

The rapid adoption of cloud, mobility, video, collaboration and big data is driving unprecedented network bandwidth and virtualization in the data center as end users increasingly expect always-on access and instantaneous application response.

Transformation in the data center has also changed the monitoring landscape tremendously. However, IT and network operators’ limited visibility into data center traffic is hampering efforts to maintain network health and ensure a high-quality user experience, not to mention fulfill regulatory and security needs

The costs and risks created by this handicap should trigger alarm bells as they extend beyond network downtime and decreased user experience to security issues, non-compliant audits and poor application performance.

Monitoring capability has quite simply not kept up with network and infrastructure upgrades. A lack of Switch Port Analyzer (SPAN) and test access point (TAP) ports feeding data to monitoring tools and the inability of existing tools to monitor traffic on both the physical and virtual networks have created blind spots. 

Changing traffic patterns

Operators’ ability to inspect packets to assess network performance and the quality of services running on the network is deteriorating. Virtualization and convergence have shifted data center traffic patterns toward ‘east-west’ traffic between applications or inter-virtual machine (VM) traffic optimized for higher speed and optimized utilization. Unfortunately, many tools are limited by throughput, hypervisor incompatibility and excessive resource utilization.

Traditional TAPs cannot see the traffic between the VMs on the same hypervisor or ‘follow’ specific VMs as they are moved from one hypervisor to another to optimize efficiency and availability. In blade servers, the problem gets more acute. Physical network tools cannot monitor traffic in each blade running multiple VMs on a hypervisor and the traffic between the blades running on the backplane.

In addition, SPAN ports capturing intra-switch traffic are known to overwhelm monitoring tools with duplicate copies of switch packets.

Meeting user expectations

“The vast majority of virtualized data centers are architected or designed with little to no thought of visibility,” says Michael Scheppke, senior director of Sales at Ixia. “This makes monitoring inter-VM or ‘east-west’ data center traffic impossible and leaves a blind spot in traffic monitoring and reporting.”  

“It also makes troubleshooting application performance issues extremely difficult since IT administrators can only see the application traffic once it hits some part of the physical network; usually the top of rack switch or a physical tap downstream from the top of rack switch,” Scheppke adds.

“To keep up with constant cycles of new application and service deployments, new use case demands, and relentless growth, network and IT organizations need deep, reliable and resilient visibility,” said Jim Frey, EMA’s vice president of Research, Network Management.

Officials at Ixia believe that next-generation cloud providers, mobility operators and enterprises must optimize visibility and control of network traffic to maintain quality of service across virtualization, application and service delivery.

For example, mobility operators’ transition from 3G to an all-IP LTE (Long Term Evolution) operating environment requires their engineers to have high visibility and real information about the mobile network to ensure performance and reliability amid a complex mix of voice and data. Today, users do not make calls and send texts only. They also want to check their email, upload images, stream videos, search the web or use cloud-based applications. 

What’s vital for IT

Clearly, end-to-end visibility is absolutely vital for any IT organization to not only control and optimize the network and the applications it delivers but also provide information to other monitoring systems deployed for security and compliance. That means an infrastructure that enables physical and virtual network, application, and security visibility. 

For this reason, Ixia’s paid US$190 million to acquire Net Optics late last year – the company’s third in the past few years. Ixia had bought another network visibility specialist, Anue Systems, in May 2012 for $145 million, and security testing specialist BreakingPoint Systems in July the same year for $160 million. 

With these acquisitions, Ixia is well placed to provide IT organizations the visibility needed to maintain a high-speed, scalable, agile and secure network that underpins the always-on user experience. Ultimately, they should make troubleshooting simpler so enterprises and service providers can easily prevent a major network fallout.