5 Key Attributes to Consider When Researching Network Performance Management Tools
by November 12, 2012

Filed under: Networking Technology, Performance Monitoring

Best NPM SolutionsNow more than ever, IT teams need to know exactly what is happening on the network, and which users and devices may be contributing to performance degradation. This level of insight is essential to gauging the impact that various activity has on the network – and ultimately to ensuring the performance of IP-based applications like cloud services, SaaS tools, VoIP, Video Conferencing, IP storage, virtual desktops and so on.

In this context, the goal of network traffic analysis tools has been to provide network and application engineers with data packet and network traffic flow information so they can monitor application and network usage, understand what applications or sessions are using what resources and pinpoint the cause of network problems. A wide range of tools exist to gather network flow data from network devices (leveraging NetFlow or SNMP), as well as through packet “sniffing” and analysis.

The problem with traditional, device-centric tools is that they don’t meet the needs of today’s complex, distributed networks.

  • SNMP, for instance, requires compatible instrumentation on network devices and can’t see beyond device health metrics to the key performance of the network.
    • Further, SNMP visibility stops at the edge of your network, providing no visibility into WAN and internet based applications like Cloud, Saas and hosted services.
  • NetFlow analysis solutions work well, but are prohibitively expensive to deploy, and they consume significant network bandwidth, making them especially problematic for use at remote sites.

Freeware style sniffers can be cheap, but the raw data feeds they provide can be challenging to interpret. Even with the proper knowledge to interpret a packet capture, the information provided is not designed to provide insight into network performance.

With the above issues and challenges in mind, here are 5 key attributes to consider when researching network performance management tools:

1) Ease of use/speed of analysis

These days, it’s vital to be able to monitor and troubleshoot the network quickly and efficiently … without needing a PhD. Look for solutions that feature a well-designed, configurable and vendor-supported dashboard that tells you what you need to know at a glance.

2) Ease of deployment

Do you need to download, distribute, install, configure and upgrade files or “agents”? Do you need to manage files or “agents” constantly as your network changes? Do you have the IT Infrastructure in place at remote locations to run these agents? Do you need to be concerned with how the various components of a traffic analysis or network performance management solution communicate, especially at remote sites?

Ease of deployment and administration is critical in this era of multitasking IT staff and remote users. Look for solutions that offer do-it-yourself, plug-and-play simplicity.

3) Level of traffic insight

Traditional network traffic analysis tools provide “port and protocol” data only – not nearly enough to monitor and troubleshoot today’s complex network traffic patterns. What is required is the ability to visualize specific applications and users and their impact on network performance – so you can spot network bottlenecks, flag the heaviest users and see what applications are using the most bandwidth.

Traditional NetFlow analysis tools do not have the granularity into the specific applications in use, which is absolutely required in any business using internal and external web applications. Web application traffic itself is not inherently good or bad, you need to understand if that activity is a critical business service like Salesforce.com or recreational activity like YouTube.com.

4) Support for remote sites

Users are working remotely more frequently these days. You need to be able to gauge the impact of their activities on the WAN. Likewise, you need to be able to monitor and troubleshoot networks at remote office locations without actually visiting them. That rules out tools that require instrumentation you don’t have at remote sites, or that otherwise aren’t easy to deploy and manage remotely.

5) Total cost of ownership/cost-effectiveness

When considering TCO for network analyzers, it’s important to look beyond purchase price. How easy is the solution to use and to manage over time? How much expensive network bandwidth does it consume? Does it require specific network hardware? Will it scale up cost-effectively as your network grows?

Do you look for these five attributes when researching a network performance management solution? If not, what do you look for?