North Carolina State University: Network Agents Plot the Path of Least Resistance

Most major U.S. universities manage enough Internet traffic to overwhelm the small squad of technicians usually on hand to oversee their networks. Like many, the three-person Web services team at North Carolina State University was becoming concerned about how to control the increasing number of Web sites, applications, and Web services, from online quizzing, testing and online course catalogues, sprouting at the school.

NSCU's information technology division, which acts as the Internet Services Provider to the school, hosts more than 50 different Web sites, including the university Web site, departmental Web sites, 95 percent of the online courses at the school, sites for service class assignments and quizzes, and sites for a number of non-profit organizations, such as the North Carolina Academy of Science and the National Humanities Center. All in all, the school records 5.5 million hits a day.

On any given day, the university might have 30 or 40 Web services residing on each server in its server farm. As these network service demands escalated, Harry Nicholos, NCSU assistant director for Unix and Web Services, began searching for a traffic management application that could maintain service availability and performance, scale with the rapid growth of the Web services, and not unduly add manpower or physical resources to the division.

The school considered solutions from IBM Corp., Cisco Inc., and F5 Inc. All three had a downside: IBM's product handled only HTTP traffic, and Nicholos wanted the flexibility to handle FTP and other TCP/IP-based traffic. Cisco's Local Director required changes to the network that the team did not want to undertake. F5's Big IP load-balancing appliance required specialized training. In the end, NCSU decided on Central Dispatch, traffic-management software from Resonate Inc.

The software is a distributed-traffic-management system that uses software agents to determine the health and availability of network servers. Depending on the traffic load on any particular server, Central Dispatch will route traffic to the most available network resource to respond to Internet requests. It also handles any type of TCP/IP–based traffic, so it fit the flexibility requirement for NCSU. The software runs on Sun Solaris, so it was compatible with NCSU's Unix environment, and installed on top of NCSU's existing network infrastructure, avoiding costs for additional hardware or network changes.

Once NCSU tackled the network availability challenge, Nicholos and his team needed a way to address the performance problems of the University's Web services. Resonate Commander Solutions, a service-level management software, allowed the team to define the performance goals and objectives they desired from the network's Web services. If those objectives were not met, the software would troubleshoot the cause.

Using software agents, the system monitors the university's infrastructure, including its Web servers, application servers, network devices and databases. These components are grouped in a logical service model, a representation of the physical infrastructure of the network as Web services are delivered. Through this service model, a diagnosis of the health of any particular Web service, all the way down to the individual component, can be made. It also takes real-time action to minimize the impact to the user or transaction while problems are identified and resolved quickly.

NCSU benefits from Commander Solutions' "inside-out/outside-in" perspective of its Web services. The software combines data on the entire Web services infrastructure with data collected from tests of simulated user transactions to better track the end user's true online experience and more quickly diagnose problems. "The quizzing service may be composed of PERL scripts connecting to Sybase … Commander Solutions tests the page and looks not only at the server, but also evaluates how the service as a whole is behaving," explains Nicholos.

If problems do occur, the diagnostic software works hand-in-hand with the traffic manager to route traffic around the problem so that NCSU's students and faculty are not affected by performance deterioration. During that time, Resonate Commander Solutions work to further diagnose the problem, fix it and alert an administrator that a problem has occurred.

NCSU is now delivering more services on the same number of servers than a "bigger-faster" strategy would require. The school's infrastructure supports more than 40 Web services and 5.5 million hits a day to its Web sites. Eight of the 16 servers used by NCSU have gone beyond their 3-year lifecycle and continue to offer Web pages at their optimal performance.

For more information, contact Harry Nicholas, NCSU assistant director for Unix and Web services at [email protected].

Featured

  • A person types on a laptop surrounded by floating digital profiles

    CrowdStrike to Acquire Identity Security Firm SGNL

    AI-based cybersecurity firm CrowdStrike has signed a definitive agreement to acquire identity security company SGNL in a deal valued at approximately $740 million.

  • Interface buttons of Generative AI tool

    Report: No Foolproof Method Exists for Detecting AI-Generated Media

    Microsoft has released a new research report warning that no single technology can reliably distinguish AI-generated content from authentic media, and that deepening reliance on any one method risks misleading the public.

  • AI word on microchip and colorful light spread

    Microsoft Unveils Maia 200 Inference Chip to Cut AI Serving Costs

    Microsoft recently introduced Maia 200, a custom-built accelerator aimed at lowering the cost of running artificial intelligence workloads at cloud scale, as major providers look to curb soaring inference expenses and lessen dependence on Nvidia graphics processors.

  • Abstract speed motion blur in vibrant colors

    3 Ed Tech Shifts that Will Define 2026

    The digital learning landscape is entering a new phase defined by rapid advances in artificial intelligence, rising expectations for the student experience, and increasing pressure to demonstrate quality and accountability in online education.