Wireless Networking at Indiana University


Published on

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Wireless Networking at Indiana University

  1. 1. Wireless Networking at Indiana University Paul Arabasz, IDC Judith Pirani, Sheep Pond Associates ECAR Case Study 5, 2002 Case Study from the EDUCAUSE Center for Applied Research
  2. 2. 4772 Walnut Street, Suite 206 Boulder, Colorado 80301 www.educause.edu/ecar/
  3. 3. Wireless Networking at Indiana University
  4. 4. EDUCAUSE is a nonprofit association whose mission is to advance higher edu- cation by promoting the intelligent use of information technology. The mission of the EDUCAUSE Center for Applied Research is to foster better decision making by conducting and disseminating research and analysis about the role and implications of information technology in higher education. ECAR will systematically address many of the challenges brought more sharply into focus by information technologies. Copyright 2002 EDUCAUSE. All rights reserved. This ECAR Research Study is proprietary and intended for use only by subscribers and those who have pur- chased this study. Reproduction, or distribution of ECAR Research Studies to those not formally affiliated with the subscribing organization, is strictly pro- hibited unless prior written permission is granted by EDUCAUSE. Requests for permission to reprint or distribute should be sent to ecar@educause.edu.
  5. 5. Wireless Networking in Higher Education Case Study 5, 2002 0 Wireless Networking at Indiana University Preface N best practices cases studies with six The EDUCAUSE Center for Applied Re- higher education institutions about their search (ECAR) produces research to pro- wireless network implementations. mote effective decisions regarding the Between March and May 2002, ECAR selection, development, deployment, man- and IDC began with a list of approximately agement, socialization, and use of informa- 150 colleges and universities that had ex- tion technology (IT) in higher education. perience implementing wireless networks. ECAR research includes research bulletins, From this list, 20 were interviewed exten- short summary analyses of key IT issues; re- sively by telephone, and six were selected search studies, in-depth applied research on for either on-site visits or extensive telephone complex and consequential technologies follow-up. On-site visits are rigorous and in- and practices; and case studies designed volve nearly two days of interviews and to exemplify important themes, trends, and meetings with the widest variety of institu- experiences in the management of IT in- tional representatives associated with—or vestments and activities. affected by—the technologies or practices ECAR has investigated the state of being investigated. wireless networking in higher education This case study was undertaken to draw and has issued “Wireless Networking in on the direct experience of others to pro- Higher Education.” This research was un- vide insights into what has—and, as appro- dertaken in three phases: priate, what hasn’t—worked in wireless N an online survey of 391 EDUCAUSE implementations. It is assumed that readers members to establish the state of wire- of the case studies will also read the main less networking in higher education report, which incorporates the findings of and to understand its implementation the case studies within the generalized con- characteristics; text of the report. N follow-up, in-depth telephone and on- ECAR wishes to thank the leadership of site interviews, covering 17 selected Indiana University for their time, assistance, institutions, with IT personnel and uni- and diligence in support of this research. We versity members who are directly in- hope readers of this ECAR case study will volved with the creation, operation, or learn from their experiences. use of wireless networks; and © 2002 EDUCAUSE. Reproduction by permission only. EDUCAUSE CENTER FOR APPLIED R ESEARCH 1
  6. 6. Wireless Networking in Higher Education Case Study 5, 2002 Introduction ing student, financial, human resources, Indiana University (IU) has eight campus procurement, facilities, research admin- locations; the largest are in Bloomington istration, instructional, library, and other (IUB), Indianapolis (IUPUI), and Fort Wayne. systems). UIS also manages the comput- The university offers four-year programs, ad- ing infrastructure supporting these infor- vanced degree programs (professional, mation systems. master’s and master’s degree equivalents, N Research and Academic Computing and doctoral), and associate’s and certificate (RAC) provides computing facilities, vi- degree programs. For the fall semester sualization facilities, support, and services 2000–2001, IU’s all-campus enrollment was for IU’s research community, including 93,775 students (graduate and undergradu- faculty researchers, clinicians, engineers, ate). With an annual operating budget of artists, and students. approximately $2.1 billion, IU employs 4,230 Within the UITS organization, the Infor- full-time and 1,901 part-time faculty and ap- mation Technology Policy Office (ITPO) co- proximately 10,500 appointed staff. Estab- ordinates the development, review, and lished in 1820 in Bloomington, Indiana, IU management of policies for a wide array of boasts 116 academic programs ranked in the IT issues. IT security issues are addressed by nation’s top 20. the Information Technology Security Office IU’s central IT organization, UITS, is (ITSO), which is closely aligned with ITPO. headed by Dr. Michael A. McRobbie, vice According to its mission statement, ITSO is president for information technology. The mandated “to provide proactive security UITS comprises four divisions: analysis, development, education, and guid- N Teaching and Learning Information Tech- ance related to Indiana University’s informa- nologies (TLIT) provides support services tion asset and information technology to faculty, students, staff, and comput- environment. The overall objective is a safe ing support professionals on the IUPUI and secure atmosphere for teaching and and IUB campuses. TLIT’s resources in- learning, research, service, and the conduct clude student computing labs and the of university business.” Knowledge Base FAQ database. A key On the university’s Bloomington campus, mandate for the division is to partner most buildings are connected to the cam- with faculty in integrating technology pus backbone using Ethernet data commu- into their teaching practices. Within the nications equipment and fiber-optic cabling UITS organization, wireless has been the (see Figure 1). Most connections are either domain of the TLIT division. Fast Ethernet (100 Mbps) or Gigabit Ethernet N Telecommunications is responsible for (1,000 Mbps). In the few, remote instances the development and evolution of IU’s where campus fiber isn’t available, buildings voice, data, and video communications are connected at a variety of speeds, de- infrastructure and devices. pending on the applications at those places. N University Information Systems (UIS) de- Speeds range from T1 (1.54 Mbps) to DS3 velops, implements, and manages the (45 Mbps), with some locations using full enterprise information systems that sup- 1-Gbps connections. port IU’s core business processes (includ- 2
  7. 7. Wireless Networking in Higher Education Case Study 5, 2002 Building #1 Main Building #2 Main Ethernet Switch Ethernet Switch 100 Mbps 100 Mbps Campus Virtual Campus Virtual LAN Switches LAN Switches 1,000 Mbps 1,000 Mbps Backbone Router Figure 1. Basic Network Architecture at Indiana 45 Mbps University’s Bloomington Campus The Internet Drivers of IU’s Wireless Planning and Funding Deployment IU’s long-term wireless plan is driven by IU’s wireless network began with a se- two basic principles that serve as the foun- ries of pilot programs conducted at its dation for the institution’s overall IT plan- Bloomington and Indianapolis campuses. ning. The first is that a wireless infrastructure The initial thrust of the wireless program was is an integral part of its strategy of estab- to provide coverage in common areas (meet- lishing itself in the top tier of institutions of ing locations and student areas) as well as public higher education. This entails the in classrooms. The most important driver “creative use and application of information behind IU’s wireless initiative was its “Infor- technology,” focusing specifically on “the mation Technology Strategic Plan,” pub- design, development, and application of in- lished in May 1998, which called for a formation technology in support of teach- converged telecom and networking infra- ing and learning, research, service, and the structure across the campus. In the context conduct of university business.” The second of the plan, wireless is seen as contributing principle is that funding for technology in to UITS’s broader strategic goal of creating general (and wireless in particular) must be a world-class infrastructure for teaching, done in a way that recognizes the full cost learning, and research. of technology, including ongoing replace- ment and support. EDUCAUSE CENTER FOR APPLIED R ESEARCH 3
  8. 8. Wireless Networking in Higher Education Case Study 5, 2002 While the first principle firmly establishes said Voss. “We essentially provide them with wireless within IU’s overall technology plan, the checklist, but they are responsible for the second has implications for the even- the procurement of the technology.” An tual scope and timing of the infrastructure example of the kinds of issues addressed by build-out. According to Brian Voss, associ- these recommendations is the appropriate- ate vice president for telecommunications, ness of frequency hopping versus direct the key to funding initiatives like wireless is sequencing in wireless infrastructure equip- to look at budgeting from a life-cycle per- ment, given IU’s security requirements. spective and to budget in the funding to While individual IU departments can maintain and upgrade the system. sponsor wireless initiatives, wireless none- “If we took the approach of using grant theless remains under the policy jurisdiction money to fund wireless, we’d probably be of UITS. Following are examples of key policy able to establish nearly ubiquitous coverage statements issued by UITS: on campus,” said Voss, “but we would still N Wireless technology is not suitable for be in a situation where we would be de- all locations and applications and is cer- pendent on ‘miracle money’ (money left over tainly not a strategic replacement for a in the budget) to maintain and upgrade the wired infrastructure. An exception is de- infrastructure as it ages and applications ployment in places where fixed wiring is become more bandwidth intensive. In gen- not an option because of building con- eral, counting on future miracles is a bad figuration, age, or location—that is, strategy for funding over the long term.” where installing traditional wiring is ei- IU’s present approach to general tech- ther not possible or not practical. nology funding shows how the process N UITS will manage all wireless hubs, ex- would likely proceed for wireless. The basic cept those that are mobile, temporary, approach to establishing technology bud- or serially connected. gets involves collaboration between UITS N All UITS-managed wireless hubs will be and individual IU departments designed to connected via the VPN-secured system, flesh out the overall cost of an initiative over unless a specific exception is granted by its projected life cycle. Once a cost has been the University Information Technology established, UITS provides departments with Policy Office. a matching grant such that the cost is N UITS will make a site visit to assist de- roughly split between individual depart- partmental staff in determining the op- ments and UITS. timal location of equipment. Where pos- sible, hubs installed by departments must Role and Policies of UITS use IU’s central, VPN-secured system. The broader role of UITS in wireless plan- Departments must contact the UITS Net- ning has been to guide departments in wire- work Operations Center to have their less technology acquisition. As Voss noted, wireless networks added to the VPN- the most important aim of these guidelines secured system. is to ensure that departments choose prod- N If equipment installed by a department ucts and solutions that adhere to IU’s pre- interferes with the wireless network vailing IT infrastructure standards. “The maintained for the university by UITS, the criteria we establish for departments is that configuration of that equipment will the wireless technology is robust, reliable, have to be changed to eliminate the con- and fits within our overall environment,” flict, or it will have to be removed. 4
  9. 9. Wireless Networking in Higher Education Case Study 5, 2002 Wireless Deployment Bloomington campus, wireless is deployed in Issues the following locations: Evolving standards, selection of areas for N Kelley School of Business deployment, and the extent and type of N Carmichael Center wireless use were among the key issues re- N Creative Arts Building quiring consideration. N Law School N Lilly Library Technology and Standards N Indiana Memorial Union Selection N Main Library When IU began piloting wireless in 2000, N Spea Library the issue of wireless standards was not of N UITS immediate importance, given that the IEEE The Kelley School of Business was one of 802.11b standard represented the only ma- the earliest adopters of wireless on the jor option at the time. However, with the Bloomington campus. One key to the success impending availability of 802.11a and of this deployment was the mandated own- 802.11g, the issue of wireless standards has ership of laptops for graduate business stu- gained prominence. The university is almost dents, which encouraged adoption by certainly moving in the direction of 802.11a, providing a proportionately larger pool of po- with the key driver being improved band- tential wireless laptop users. Another was the width. However, as Voss explained, its gen- high level of importance graduate business eral “newness” acts as a gating factor. “We students attached to mobility—specifically the will deploy 802.11a when we feel it has ability to move from room to room without matured enough to be reliable, and the costs having to unplug. The fact that many (if not are such that we can reasonably replace the most) graduate business students choose to equipment on a standard three- or four-year use their wireless connections even when life cycle.” ample wired connections are available signi- One factor that may accelerate 802.11a fies how embedded wireless has become in deployment is an interest within certain de- their everyday collaboration. partments in moving more quickly toward Deployment of wireless in common areas its adoption. To ensure standardization, UITS was an especially important part of IU’s ini- acknowledges that it may be compelled to tial roll-out plan. The two most important identify an 802.11a-compliant product ear- such areas were the Indiana Memorial Union lier than it had originally planned. (where wireless is now fully deployed) and the main library. The goal of the library Scope and Focus of Wireless deployment was to provide greater mobility Deployment for students and researchers. In the first phase of deployment, IU de- ployed wireless in “islands” across its cam- Primary Applications puses, largely because of its goal of For the university as a whole, Web brows- providing the most coverage with the least ing and messaging represent the dominant amount of funding. The locations of these wireless applications. While there are relatively islands—a mixture of academic classrooms, few examples of departments using wireless meeting places, and student gathering ar- for highly specific applications, the Computer eas—were influenced by individual depart- Science Department has been relatively aggres- ments, which also provided funding. On IU’s sive in using wireless for collaboration and ex- EDUCAUSE CENTER FOR APPLIED R ESEARCH 5
  10. 10. Wireless Networking in Higher Education Case Study 5, 2002 panding access to its server resources. Simi- truders. While UITS controls all aspects of the larly, the UITS organization uses wireless in the wired network (including security), the free- course of everyday departmental activities, dom of departments to add wireless infra- explained Voss, a frequent user of wireless him- structure establishes a potential point of self. “Wireless has been an indispensable tool vulnerability in the area of network security. in department meetings because it allows our UITS addressed this by establishing a set of staff to access and alter reports in real time,” policies and standards, an important element he said. “This increases the overall productiv- of which is the requirement that departments ity of the department.” plug into the VPN. Architecture and Perceived Benefits of Security Profile Wireless IU’s wireless implementation employs a Brian Voss sees the primary benefits of mix of Lucent 500 and 1,000 and Cisco a wireless infrastructure as wider and more Aironet wireless access points. The location convenient access to IU’s network re- of each access point was determined by a sources, beyond the range of wired bound- series of site studies conducted by UITS. Dur- aries. “IU is already a top ‘most-wired’ ing its site surveys, UITS determined the lo- institution, and we’d like to become a top cation and number of access points required ‘most-wireless’ as well,” he said. “We see to provide a given area with a throughput the complementary combination of wired of at least 2 Mbps. For sites with a height- and wireless infrastructures as benefiting ened probability of airwave congestion, UITS all constituencies at IU.” used dual-transmitter access points to While Voss sees such metrics as network double the number of simultaneous users. usage as a gauge of success, he believes that IU’s wireless architecture places all users on the real value of wireless is seen in more a single subnet, with a net mask that sup- nebulous—and far-reaching—measures. ports 1,024 users. When simultaneous us- “UITS looks at the value of technology in age rates begin to exceed network capacity terms of its potential for facilitating the (not expected until mid-2003), UITS expects teaching, learning, research, and service to split its wireless network into two domains missions of the university. It’s less a ques- of 1,024 users each. tion of metrics than of providing a fertile IU’s wireless network is secured through environment via technology.” a VPN that provides both end-user authen- tication (via a user name and password) and Lessons Learned encryption. The UITS organization manages Although still in the early stages of its and maintains the VPN environment, which wireless deployment, IU has learned several is designed to allow departments coming key lessons, the most important of which on line with wireless to simply “plug in.” concerning how wireless should be intro- The VPN solution has thus far delivered ac- duced. Brad Wheeler, associate dean in ceptable security and prevented significant UITS’s Teaching and Learning Information security problems. Technologies Division, sees the value of rapid IU’s focus on security reflects the deployment (over an incremental approach) institution’s relatively high IT profile within the as an important lesson of IU’s experience. higher education community, a status that “With technology initiatives like wireless, it’s has made it a frequent target of network in- better to embrace it on a global scale,” said 6
  11. 11. Wireless Networking in Higher Education Case Study 5, 2002 Wheeler, “changing not just the technology “While we’re still trying to work it out, we infrastructure but the processes and fund- see two alternatives—technical solutions ing priorities as well.” and behavioral solutions. I see technical so- Brian Voss concurred and pointed to the lutions as inferior because it turns into a ‘ra- need to match expectations with capabili- dar and radar detector’ scenario. At the end ties—as well as with the funding required of the day, behavioral solutions are the only to back them up. “Because Indiana is a answer.” leader in IT, there was an expectation that we were going to build a wireless infrastruc- The Future of Wireless ture that would enable people to access at Indiana University from anywhere on campus,” noted Voss. Going forward, IU expects its wireless “This shows we needed to better set expec- coverage to grow steadily and quickly. IU still tations or define an adequate funding plans to build increasingly larger islands of stream to enable ubiquitous deployment.” connectivity, which will eventually deliver IU’s wireless experience also uncovered campus-wide coverage. But, in step with and debunked widely held misconceptions Brad Wheeler’s comments, the intention is about the respective roles of the wireless and to do that over a fairly short period of time. wired infrastructure, and the funding impli- The effort began in the summer of 2002 and cations that sprang from these misconcep- will proceed rapidly in stages of deployment, tions. The first of these was the simplistic culminating in a campus-wide solution for assumption that the wireless infrastructure IU’s two core campuses, in Bloomington and would ultimately replace large components Indianapolis, by fall 2003. of the wired infrastructure, enabling large The key, said Voss, has been to find a portions of funding for the wired network to funding model that can make broad-based be channeled to wireless initiatives. “We wireless deployment a reality. “We started to should have understood from the outset that think out of the box [on wireless funding],” we were talking about an augmentation de- he said. “We believe, now, we can reallocate ployment, not a substitution deployment,” both one-time and ongoing funds from other explained Voss. “It would have given us a services and savings exercises—particularly more realistic funding plan earlier on.” The low-speed campus network access via mo- second (and a related) assumption that was dems, which is rapidly giving way to high- proved wrong was that reductions in the cost speed, user-acquired access in the form of of maintaining the wired infrastructure would cable modems and DSL services.” This tran- free up funds from the existing budget to sition means that IU will be trading its invest- support a broader wireless deployment. ment in low-speed modem access for As IU expands its embrace of wireless, high-speed wireless access on the campuses. one of the challenges it faces is to adapt The shift corresponds with users’ migration, classroom testing practices—which often brought about by their own demands for permit laptop use—to a wireless comput- developing their home connections. ing environment. The university’s short-term How will IU’s continued deployment be solution has been to require students to re- carried out? Voss said UITS will quickly evalu- move their wireless cards during a test. How- ate a number of alternatives. “We have an ever, with more and more laptops shipping opportunity to do the best thing we can. with wireless built into the motherboard, IU’s [That may be] a continuation of our self- Wheeler sees the challenge escalating. deployment/management strategy, such as EDUCAUSE CENTER FOR APPLIED R ESEARCH 7
  12. 12. Wireless Networking in Higher Education Case Study 5, 2002 we use on our wired network, or going with cost-effective, though it allows only a lim- an outsourced installation of access points ited amount of bandwidth,” Voss explained. and integration of that environment with “With a limited number of machines— our current network and security environ- which describes the current situation— ment.” He added, “Or perhaps something 802.11b will be fine. But when more and in between. Whatever it turns out to be, more people are looking for wireless con- we’ll be determining that very quickly over nectivity down the road, or when the next the next 60–90 days.” generation of high-bandwidth applications On the technology front, IU continues comes along, we’ll be straining that to evaluate IEEE 802.11a, 802.11g, and even network’s ability to deliver adequate perfor- G3. But in the near term, as it moves to de- mance. Fortunately, it looks like the newer ploy wireless rapidly across its campuses, equipment will have fairly straightforward UITS believes it will follow the option of de- upgrade paths, at least within the 802.11x ploying 802.11b immediately. “802.11b cer- technology. So we believe we’ll be fine.” tainly works, is well established, and is 8