Search /
Docfinder:
Advanced search  |  Help  |  Site map
RESEARCH CENTERS
SITE RESOURCES
Click for Layer 8! No, really, click NOW!
Networking for Small Business
TODAY'S NEWS
Valentine's Day Patch Tuesday: Microsoft to issue 9 patches, 4 critical
Mobile World Congress sneak peek: Quad-core smartphones, Ice Cream Sandwich & more
Microsoft details 'Windows on ARM' program
March debut of 'iPad 3' a sure bet, says analyst
FBI unbolts Steve Jobs 1991 investigation file
Cisco boosted profit, sales in Q2 while cutting costs
Macs take on the enterprise
Four crazy tech ideas from Google's Solve for X project
Obama 2012 campaign playlist revealed courtesy of Spotify
Oracle buying Taleo for US$1.9 billion in direct hit at SAP
Amazon attacks Apple: You get 3 Kindle products for price of iPad 2
Pre-rendered pages highlight latest Google Chrome release
Microsoft exec: Lync-Skype integration a 'compelling opportunity'
The future of hypervisors
/

Review: NetQoS's SuperAgent 2.0

Change the way you diagnose performance problems.

Related linksToday's breaking news
Send to a friendFeedback
Related linksToday's breaking news
Send to a friendFeedback



Diagnosing difficult network and application-layer problems is often an exercise in frustration and finger-pointing. The network staff is certain the packets are flowing fine. The database administrators swear Oracle is crunching out thousands of transactions per second. And the server team says they're only seeing a 10% CPU load. But the help desk phones are ringing, and users are getting perturbed. Sound familiar? Wish it didn't? Then take a look at NetQoS's SuperAgent 2.0.

NetQoS provides network and enterprise application managers with a powerful tool to pinpoint problems. Put SuperAgent on the case, and you can forget about the finger-pointing.

NetQoS's SuperAgent 2.0 is a network and application response-time analyzer that's completely passive - no endpoints, server applets or Internet Control Messaging Protocol (ICMP) ping packets needed. In fact, SuperAgent doesn't put any additional traffic on the network. Based on a Dell PowerEdge server and Windows 2000, SuperAgent uses sophisticated statistical analysis techniques to gather end-to-end performance numbers on any TCP-based application. Results are stored in a SQL database and displayed via a Web-based interface. Reports and alarms are delivered by e-mail or through an existing management system. SuperAgent is available in two models: a stand-alone probe, which we tested, and a multiunit configuration that aggregates the data collection and reports functionality from multiple collectors.

Targeted at the enterprise, SuperAgent lands squarely in the gray area between a full-blown protocol analyzer and an active application measurement package with client-side agents and server applets. Other products, such as Mercury Interactive's Topaz, provide some of the same features, but SuperAgent is unique in its passive approach to TCP-based network and application analysis.

On the case

We deployed SuperAgent in a harsh environment - our enterprise server network. Our bandwidth utilization averages 55M bit/sec and the applications on this network segment are critical to our daily operations: SAP R/3, PeopleSoft, our e-mail server farm, Web server farm and primary FTP site.

Initial setup of SuperAgent is straightforward and a "first-time setup" wizard is provided to simplify the process. To begin data collection and reporting, several configuration parameters need to be defined. NetQoS uses the concept of Client Regions to distinguish between different user groups. A client region is a range of IP addresses and can be configured as a Class A, B, C or D network.


How we did it
NetResults
Archive of Network World reviews
Subscribe to the Product Review newsletter


We configured SuperAgent to use our two Class B network addresses: 152.3.0.0 and 152.16.0.0. A word of warning: We discovered that using global addresses such as our two Class Bs, while simplifying the initial setup process, made later configuration more complex. Once we told SuperAgent that our client regions were 152.3.0.0 and 152.16.0.0, the application would not let us go back and define a subnet contained within these Class B networks. To identify 152.3.2.0 as the "Marketing" client region we had to delete the global definitions. Network administrators with responsibility for a range of addresses might find this cumbersome. We would prefer to define global regions while retaining the ability to specify individual subnets as unique client areas.

After connecting SuperAgent to a mirrored port and verifying network connectivity, we used the autodiscover feature to analyze the traffic flowing across our network. Autodiscover looks at the TCP port numbers and builds a table of servers and applications. The autodiscover process was very useful and saved us from having to manually configure each application and server that we wished to monitor. Unfortunately, the current release of SuperAgent does not perform reverse DNS lookups, so IP address and not domain name identified the servers. We found this to be inconvenient, particularly when several servers are grouped under the same domain name in a server farm. NetQoS officials said reverse DNS functionality will be provided in release 3.0, due out late this summer.

Rapid reports

SuperAgent provides users with a wealth of information. The graphs generated by SuperAgent were very useful - detailed, easy to read and to the point. The amount of statistics collected by SuperAgent is enormous, and NetQoS gets high marks for not burdening network executives with information overload. Each graph contains just the right amount of detail. If an in-depth look at data is desired, detailed descriptions are only a click away.

Graphs available include:

  • Response Time Summary: The time elapsed from when a client requests data to receipt of the last packet in the response.

  • Traffic: Reports on the volume of bytes, packets, bit and packet rates.

  • Sessions: Session count, refused session count and session times.

  • Components: Connection set-up time, server response time, data transfer, retransmission delay and network round-trip time.

  • Quality of service: Rate per user, number and IP address of users and loss statistics.

  • Trends: Daily, weekly and monthly.

    We were particularly impressed by SuperAgent's ability to calculate network round-trip time (RTT) without using ICMP ping packets. SuperAgent does this by taking advantage of the sequence numbers used in TCP sessions. The first sequence number is noted and time-stamped, and each subsequent response from the server is recorded. Network round-trip time is calculated by using the difference between the time stamp of the first client request and the subsequent acknowledgement packet from the server. Application response time is not a factor. We verified the validity of SuperAgent's RTT calculations by running a parallel ping process and comparing the results.

    The graphs are available through the Web-based interface, or they can be sent to users via e-mail. We encountered two problems with the e-mail delivery of the graphs. The most serious problem was related to the lack of reverse DNS lookups. The reports from the SuperAgent probe were e-mailed with a return address of SuperAgent@152.3.x.x instead of SuperAgent@netqos.netcom.duke.edu. Because our e-mail systems reject all mail from unknown domains, the reports were never received. NetQoS responded immediately to our request for help and made a patch available on its Web site within 24 hours. We installed the patch, and the problem was resolved.

    Our second problem was that Lotus Notes doesn't provide support for the PRG graphic format used by SuperAgent. NetQoS said JPG support would be provided in the next release.

    Installation and documentation

    SuperAgent 2.0 comes preinstalled on a Dell PowerEdge server running Win 2000. Installation is straightforward. Simply locate SuperAgent as close as possible to the servers being monitored and enable port mirroring on the probe's Ethernet connection. That's it - with a Web-based interface there's no client application to install.

    The documentation for SuperAgent 2.0 is top-notch. Hyperlinks on the SuperAgent home page provide ready access to NetQoS's frequently asked questions, release notes, downloads and documentation. Two printed manuals are supplied: a concise SuperAgent user's guide and a comprehensive SuperAgent administrator's guide. We found both to be excellent.

    SuperAgent 2.0's combination of functionality, reliability and superb reports make it a definite "must have."


    How we did it

    We installed the SuperAgent 2.0 probe on our enterprise application network segment. The probe's Ethernet adapter was connected to a mirrored port on the Cisco 3548 that handles all traffic destined for our SAP R/3, PeopleSoft, e-mail, FTP and Web servers. Bandwidth utilization on this segment averages 55M bit/sec.

    We let SuperAgent collect data for two weeks to test reliability, functionality and e-mail notification. The probe maintained a consistent response time during the evaluation process despite rapidly-increasing growth in the database and never crashed during our test period.



    SuperAgent 2.0
    4.25
    Rating
    Company: NetQos, (877) 835-9575, www.netqos.com Cost: $28,500 for stand-alone probe; $45,000 for management console and $9,500 for each collector. Pros: Easy to install; superb statistics; rock-solid reliability; great documentation; excellent reporting facilities. Cons: Likely to remain a Windows-centric tool; high cost.  
    SuperAgent
    Features 40% 4  
    Ease of use 35% 4  
    Reliability 10% 5  
    Installation 10% 5  
    Documentation 5% 5  
    TOTAL SCORE 4.25  
    Individual category scores are based on a scale of 1 to 5. Percentages are the weight given each category in determining the total score. Scoring Key: 5: Exceptional showing in this category. Defines the standard of excellence; 4: Very good showing. Although there may be room for improvement, this product was much better than the average; 3: Average showing in this category. Product was neither especially good nor exceptionally bad; 2: Below average. Lacked some features or lower performance than other products or than expected; 1: Consistently subpar, or lacking features being reviewed.

    Currier is the ECI Director of Data and Telecommunications at Duke University in Durham, N.C., and the 1998 Grand Prize winner in the Excellence in Campus Networking competition sponsored by CAUSE, a user group for computer professionals in higher education. He can be reached at robert.currier@duke.edu.


    NW Test Alliance

    Global Test Alliance

    Currier is also a member of the Network World Global Test Alliance, a cooperative of the premier reviewers in the network industry, each bringing to bear years of practical experience on every review. For more Test Alliance information, including what it takes to become a member, go to www.nwfusion.com/alliance.

  • RELATED LINKS


    NWFusion offers more than 40 FREE technology-specific email newsletters in key network technology areas such as NSM, VPNs, Convergence, Security and more.
    Click here to sign up!
    New Event - WANs: Optimizing Your Network Now.
    Hear from the experts about the innovations that are already starting to shake up the WAN world. Free Network World Technology Tour and Expo in Dallas, San Francisco, Washington DC, and New York.
    Attend FREE
    Your FREE Network World subscription will also include breaking news and information on wireless, storage, infrastructure, carriers and SPs, enterprise applications, videoconferencing, plus product reviews, technology insiders, management surveys and technology updates - GET IT NOW.