february 2008
     
   
 
2007 Revenues
Rise by 47 Percent
Fueled by record GigaStor™ sales, expanded analysis offerings, and increased demand for 10 Gb analysis solutions, total company revenue increased by 47 percent in 2007. This marks the thirteenth consecutive year of double-digit growth. The surge in company sales was driven by significant competitive displacements and demand for retrospective analysis, reporting, and infrastructure management solutions.
     
   
  Using NetFlow to Obtain
Cost Effective Visibility
Gaining visibility into routed traffic across every port on a global network can be a large investment using probes alone. Partnering NetFlow with an analyzer like Network Instruments® Observer® is a cost-effective way to extend overall network visibility and conduct in-depth packet analysis.

NetFlow is a Cisco-developed “push” technology embedded on many routers and switches. NetFlow statistics are captured and processed by the Layer 3 devices and sent periodically to designated collectors like Observer. It can operate as a first-line diagnostic monitor by detecting abnormal activity as it occurs. For example, NetFlow could detect an end-user station consuming excessive bandwidth. Because the technology is commonly included on enterprise infrastructure, no additional investment beyond the flow collectors is needed to achieve enterprise-wide visibility.

Although NetFlow provides high-level visibility, the only way to perform root-cause analysis is by analyzing the packets to identify the source of network problems. Diagnostic methods such as application response time analysis and connection dynamics require a complete packet stream, which is not available with NetFlow. But, NetFlow can be easily coupled with Observer to provide both high-level visibility and in-depth problem investigation through packet-level Expert Analysis.

With Network Instruments Advanced Expert Probe as a NetFlow collector, an organization can monitor high-level performance metrics, including real-time statistics like Top Talkers, bandwidth usage, and long-term network trends. The engineer could then use Network Instruments Observer and run Expert Analysis to further investigate.

Using both NetFlow metrics and Observer analysis is especially effective in detecting and resolving application performance issues. NetFlow identifies which applications are active on the network and provides usage statistics and trends on those applications. The engineer can then drill down to analyze the application’s performance with Observer.

In conclusion, using Network Instruments probes to collect both packets and NetFlow metrics provides a complete network picture. Observer harnesses the high-level visibility of NetFlow, and provides in-depth packet analysis to resolve problems allowing you to further maximize your existing infrastructure investment.

     
   
  Selecting and Comparing
Captures from Multiple Instances
This tip comes in response to customers who asked our technical support engineers how to capture and compare data from local and remote probes for MultiHop Analysis. In MultiHop Analysis, Observer automatically correlates the timing of packets between local and remote packet captures.

By comparing multiple captures from different ends of the network, you can determine:

  • If the packets reached their destination
  • Delay issues at each hop
  • The approximate location of a failure

The following steps outline the process for two probes. The same steps would be used for MultiHop Analysis involving captures from more than two probes.

  1. From the main Observer console screen, select Capture from the main menu. From the Capture drop-down menu choose Packet Capture on Multiple Instances.

  2. From the Packet Capture on Multiple Instances Settings window, select the two instances you’d like to start a capture on. You can also configure a filter for specific traffic by clicking on the Set Filter for Selected Instance button. Press Start to begin the packet captures.



  3. The main Observer screen will display the captures of the selected instances. The Multiple Instance Packet Capture box will appear above the instances. When you have collected enough data, check both the Transfer and Save Packet Captures box and the Start MultiHop Analysis box. Click Stop to transfer the remote capture to your local Observer console and begin MultiHop Analysis.

  4. The MultiHop Analysis wizard leads you through the correlation process. In the MultiHop Analysis Settings window, your instances should already be selected. You can add another instance, or press OK to begin MultiHop Analysis.

 
     
 
 


Do you consider yourself to be a packet head? Now you can prove it. Every month we’ll ask a question related to networking.

Submit the right answer and you’ll be entered to win a Network Instruments polo shirt. The winner and answer will be announced in the March edition of the Eye.

What is Cisco’s default OSPF cost for a T1 link?

Send your answer.

10 Gb Adoption
and Network Analysis
Learn about considerations to make when monitoring 10 Gb

Flow Technologies White Paper
Use flow technologies to extend your network visibility.

IT Roadmap Show
March 4
Denver, CO

NIU in Chicago, IL
March 10 – 13

NIU in Brasted, Kent, UK
April 15 – 17