Connect Banner
current issue

Wireless at 26th Street

A Pilot Program
at an NYU Residence Hall

Carlo Cernivani

Nearly six years ago, NYU's Information Technology Services (ITS) launched a pilot wireless network that eventually evolved into the NYURoam service widely available at the University today. Around that time, WiFi technologies were just beginning to get a foothold in schools, homes, and businesses. The demand for wireless technologies has risen steadily since.

In the network's first month of limited availability, slightly more than 2,000 members of the NYU community used NYURoam, a number that heavily reflected the early adoption of WiFi technology by NYU's School of Law. Since that time, the number of individuals accessing NYU-NET wirelessly per semester has increased dramatically. In the fall 2008 semester, more than 26,000 community members accessed NYU-NET and Internet resources via NYURoam. This figure does not include guests (non-NYU-affiliated individuals), who are granted temporary access while visiting the University.

In order to address both growing demands for and advances in wireless technology, NYURoam's coverage has been expanded steadily, and its radio components have undergone two complete upgrades, from IEEE 802.11b-only support to 802.11g support, and the subsequent addition of 802.11a radios to all of our deployed equipment. Further, ITS continues to evaluate, on an ongoing basis, strategies for both meeting increased demand and taking advantage of the service improvement opportunities offered by technology advances.

Currently, all of NYURoam's wireless access points (WiFi network hardware that receives and transmits radio traffic) have been configured and managed as autonomous devices. That is, each device is its own entity, loaded with its own unique configuration, and tailored to the environment in which it will be delivering service. This operational model has long been accepted as a best practice, and is a solid service delivery model. It will soon be wanting, however, in the face of burgeoning demands for wireless service from the NYU community, industry trends, and the need to better monitor and control the WiFi environment we provide.

The current accepted best strategy for deploying and managing large-scale enterprise wireless networks like NYURoam is by means of a centrally controlled and monitored "unified" network.

ITS began evaluating this new approach to WiFi service delivery late in 2007 and, as part of this process, deployed a pilot network of this type at NYU's 26th Street residence hall in summer 2008. The new network would provide wireless coverage throughout the entire building, enabling ITS to use an array of metrics to determine whether the infrastructure and operational model on which it was based could meet NYU's requirements.

A New Wireless Standard

Three key components were deployed as part of the new network's architecture. The first was a set of some 75 IEEE 802.11n pre-standard access points (APs) that were installed in the building. These devices send and receive radio transmissions to and from laptops and other radio clients.

Each of these new APs has six radios (three for each supported radio frequency, 2.4 GHz and 5 GHz), a hardware enhancement from our existing APs, which had two radios (one for 802.11g 2.4 GHz and 802.11a 5 GHz). The four additional radios are required to support the emerging 802.11n standard.[1]

The 802.11n standard has been heralded by many as a substantial leap forward for wireless technology. This is because of 802.11n's high level of reliability and bandwidth delivery. Estimates of what the maximum bandwidth deliverable from an 802.11n AP fluctuate greatly depending on whom you ask. However, a generally accepted figure is approximately 200 Mbps per access point, to be shared by clients connecting through that particular AP. This is close to four times the bandwidth delivered by our current wireless access points.

Our 802.11n access points are configured to operate as "lightweight" APs, meaning that they are controlled by the second component of this new network architecture — a wireless LAN controller or Wireless Services Module (WiSM).

Managing An Integrated Network

WiSMs perform an array of functions; they are the workhorse of the unified WiFi network architecture. In the broadest terms, a WiSM configures, monitors, and fully manages all of the access points under its control. It establishes the security policies for each of its WiFi networks, performs extensive real-time Radio Frequency (RF) monitoring, and, as is the case with our test deployment, makes real-time decisions on what operational characteristics (e.g., transmission power level, radio channel) to adjust as it addresses interference issues, radio congestion, active security threats or attacks, "holes" in the service coverage areas, and more.

A single WiSM can support as many as 300 APs simultaneously. In our NYU-NET environment, these controllers would be installed in the routers positioned out on the network. Further, access points can be instructed to reroute their data traffic to a different WiSM located elsewhere on NYU-NET should its primary WiSM fail.

Our new network architecture's third component is the Wireless Control System (WCS), a powerful software package that enables many centralized functions. It can manage all the WiSMs on our network and, in turn, all the APs in operation, and has many other capabilities. As with the WiSMs, the WCS has too many features to list here. It provides a central point for WiFi management, policy enforcement, troubleshooting, LAN design, security monitoring, intrusion detection, and more. The extensive quantity of historic and real-time data generated by WiFi activity is gathered by the WiSMs and APs and is used by the WCS, enabling complete visibility into the network from many management and monitoring perspectives.

Figures 1 through 4 provide just a few illustrations of the breadth and depth of data and monitoring information made available at one's fingertips through use of the WCS.

The WCS has numerous and varied analytical tools that are invaluable to managing an integrated network that needs to deliver the best possible wireless environment for today's applications and the mobility-oriented applications of the future. Having a reliable, secure, and flexible wireless network that addresses the various mission-critical elements that make up a complex institution such as NYU is not a luxury but a necessity. A unified wireless network architecture helps lay the foundation for the University's IT future.

Figure 1. Usage data can be monitored and displayed in a number of ways. Here we see the number of clients simultaneously connected through one of our WiSMs over a four-week period.

Figure 2. The WCS can be loaded with floor maps of coverage areas which can then be used to predict signal and coverage delivery to that floor relative to the access points deployed and the obstructions (i.e., sheetrock walls, cinder block walls, doors, etc.) that are in the environment. (Image courtesy Cisco Systems, Inc.)

Figure 3. The graph displays the downstream (data transmitted to our wireless clients) and upstream (data transmitted from our wireless clients) through a specific WiSM controller over a four-week period. During this period, the downstream bandwidth peaked at almost 30 Mbps and upstream at over 12 Mbps.

Figure 4. These graphs represent the percentage of radio frequency utilization by a specific access point for both the 2.4 GHz (802.11b/g/n) & 5 GHz (802.11a/n) bands.

FOOTNOTES

[1] Although the IEEE standards body has yet to ratify the current draft of 802.11n, the networking industry is moving forward by deploying the current "accepted" 802.11n pre-standard, confident that any modifications to the pre-standard would be addressed by simple software or firmware updates.

Author Biography

Carlo Cernivani is a Senior Project Manager for Planning and Implementation within ITS' Communications & Computing Services.