wireless networking at indiana university

12
Wireless Networking at Indiana University Case Study from the EDUCAUSE Center for Applied Research Paul Arabasz, IDC Judith Pirani, Sheep Pond Associates ECAR Case Study 5, 2002

Upload: networksguy

Post on 22-Dec-2014

301 views

Category:

Documents


0 download

DESCRIPTION

 

TRANSCRIPT

Page 1: Wireless Networking at Indiana University

Wireless Networkingat Indiana University

Case Study from theEDUCAUSE Center for Applied Research

Paul Arabasz, IDCJudith Pirani, Sheep Pond Associates

ECAR Case Study 5, 2002

Page 2: Wireless Networking at Indiana University

4772 Walnut Street, Suite 206Boulder, Colorado 80301www.educause.edu/ecar/

Page 3: Wireless Networking at Indiana University

Wireless Networkingat Indiana University

Page 4: Wireless Networking at Indiana University

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 betterdecision making by conducting and disseminating research and analysis aboutthe role and implications of information technology in higher education. ECARwill systematically address many of the challenges brought more sharply intofocus by information technologies.

Copyright 2002 EDUCAUSE. All rights reserved. This ECAR Research Study isproprietary and intended for use only by subscribers and those who have pur-chased this study. Reproduction, or distribution of ECAR Research Studies tothose not formally affiliated with the subscribing organization, is strictly pro-hibited unless prior written permission is granted by EDUCAUSE. Requests forpermission to reprint or distribute should be sent to [email protected].

Page 5: Wireless Networking at Indiana University

EDUCAUSE CENTER FOR APPLIED RESEARCH 1

Wireless Networking in Higher Education Case Study 5, 2002

© 2002 EDUCAUSE. Reproduction by permission only.

0Wireless Networkingat Indiana University

PrefaceThe EDUCAUSE Center for Applied Re-

search (ECAR) produces research to pro-mote effective decisions regarding theselection, development, deployment, man-agement, socialization, and use of informa-tion technology (IT) in higher education.ECAR research includes research bulletins,short summary analyses of key IT issues; re-search studies, in-depth applied research oncomplex and consequential technologiesand practices; and case studies designedto exemplify important themes, trends, andexperiences in the management of IT in-vestments and activities.

ECAR has investigated the state ofwireless networking in higher educationand has issued “Wireless Networking inHigher Education.” This research was un-dertaken in three phases:� an online survey of 391 EDUCAUSE

members to establish the state of wire-less networking in higher educationand to understand its implementationcharacteristics;

� follow-up, in-depth telephone and on-site interviews, covering 17 selectedinstitutions, with IT personnel and uni-versity members who are directly in-volved with the creation, operation, oruse of wireless networks; and

� best practices cases studies with sixhigher education institutions about theirwireless network implementations.Between March and May 2002, ECAR

and IDC began with a list of approximately150 colleges and universities that had ex-perience implementing wireless networks.From this list, 20 were interviewed exten-sively by telephone, and six were selectedfor either on-site visits or extensive telephonefollow-up. On-site visits are rigorous and in-volve nearly two days of interviews andmeetings with the widest variety of institu-tional representatives associated with—oraffected by—the technologies or practicesbeing investigated.

This case study was undertaken to drawon the direct experience of others to pro-vide insights into what has—and, as appro-priate, what hasn’t—worked in wirelessimplementations. It is assumed that readersof the case studies will also read the mainreport, which incorporates the findings ofthe case studies within the generalized con-text of the report.

ECAR wishes to thank the leadership ofIndiana University for their time, assistance,and diligence in support of this research. Wehope readers of this ECAR case study willlearn from their experiences.

Page 6: Wireless Networking at Indiana University

2

Wireless Networking in Higher Education Case Study 5, 2002

IntroductionIndiana University (IU) has eight campus

locations; the largest are in Bloomington(IUB), Indianapolis (IUPUI), and Fort Wayne.The university offers four-year programs, ad-vanced degree programs (professional,master’s and master’s degree equivalents,and doctoral), and associate’s and certificatedegree programs. For the fall semester2000–2001, IU’s all-campus enrollment was93,775 students (graduate and undergradu-ate). With an annual operating budget ofapproximately $2.1 billion, IU employs 4,230full-time and 1,901 part-time faculty and ap-proximately 10,500 appointed staff. Estab-lished in 1820 in Bloomington, Indiana, IUboasts 116 academic programs ranked in thenation’s top 20.

IU’s central IT organization, UITS, isheaded by Dr. Michael A. McRobbie, vicepresident for information technology. TheUITS comprises four divisions:� Teaching and Learning Information Tech-

nologies (TLIT) provides support servicesto faculty, students, staff, and comput-ing support professionals on the IUPUIand IUB campuses. TLIT’s resources in-clude student computing labs and theKnowledge Base FAQ database. A keymandate for the division is to partnerwith faculty in integrating technologyinto their teaching practices. Within theUITS organization, wireless has been thedomain of the TLIT division.

� Telecommunications is responsible forthe development and evolution of IU’svoice, data, and video communicationsinfrastructure and devices.

� University Information Systems (UIS) de-velops, implements, and manages theenterprise information systems that sup-port IU’s core business processes (includ-

ing student, financial, human resources,procurement, facilities, research admin-istration, instructional, library, and othersystems). UIS also manages the comput-ing infrastructure supporting these infor-mation systems.

� Research and Academic Computing(RAC) provides computing facilities, vi-sualization facilities, support, and servicesfor IU’s research community, includingfaculty researchers, clinicians, engineers,artists, and students.Within the UITS organization, the Infor-

mation Technology Policy Office (ITPO) co-ordinates the development, review, andmanagement of policies for a wide array ofIT issues. IT security issues are addressed bythe Information Technology Security Office(ITSO), which is closely aligned with ITPO.According to its mission statement, ITSO ismandated “to provide proactive securityanalysis, development, education, and guid-ance related to Indiana University’s informa-tion asset and information technologyenvironment. The overall objective is a safeand secure atmosphere for teaching andlearning, research, service, and the conductof university business.”

On the university’s Bloomington campus,most buildings are connected to the cam-pus backbone using Ethernet data commu-nications equipment and fiber-optic cabling(see Figure 1). Most connections are eitherFast Ethernet (100 Mbps) or Gigabit Ethernet(1,000 Mbps). In the few, remote instanceswhere campus fiber isn’t available, buildingsare connected at a variety of speeds, de-pending on the applications at those places.Speeds range from T1 (1.54 Mbps) to DS3(45 Mbps), with some locations using full1-Gbps connections.

Page 7: Wireless Networking at Indiana University

EDUCAUSE CENTER FOR APPLIED RESEARCH 3

Wireless Networking in Higher Education Case Study 5, 2002

Building #1 Main

Ethernet Switch

100 Mbps

Campus Virtual

LAN Switches

Backbone

Router

1,000 Mbps

The Internet

45 Mbps

Building #2 Main

Ethernet Switch

100 Mbps

Campus Virtual

LAN Switches

1,000 Mbps

Drivers of IU’s WirelessDeployment

IU’s wireless network began with a se-ries of pilot programs conducted at itsBloomington and Indianapolis campuses.The initial thrust of the wireless program wasto provide coverage in common areas (meet-ing locations and student areas) as well asin classrooms. The most important driverbehind IU’s wireless initiative was its “Infor-mation Technology Strategic Plan,” pub-lished in May 1998, which called for aconverged telecom and networking infra-structure across the campus. In the contextof the plan, wireless is seen as contributingto UITS’s broader strategic goal of creatinga world-class infrastructure for teaching,learning, and research.

Planning and FundingIU’s long-term wireless plan is driven by

two basic principles that serve as the foun-dation for the institution’s overall IT plan-ning. The first is that a wireless infrastructureis an integral part of its strategy of estab-lishing itself in the top tier of institutions ofpublic higher education. This entails the“creative use and application of informationtechnology,” focusing specifically on “thedesign, development, and application of in-formation technology in support of teach-ing and learning, research, service, and theconduct of university business.” The secondprinciple is that funding for technology ingeneral (and wireless in particular) must bedone in a way that recognizes the full costof technology, including ongoing replace-ment and support.

Figure 1. Basic NetworkArchitecture at IndianaUniversity’s BloomingtonCampus

Page 8: Wireless Networking at Indiana University

4

Wireless Networking in Higher Education Case Study 5, 2002

While the first principle firmly establisheswireless within IU’s overall technology plan,the second has implications for the even-tual scope and timing of the infrastructurebuild-out. According to Brian Voss, associ-ate vice president for telecommunications,the key to funding initiatives like wireless isto look at budgeting from a life-cycle per-spective and to budget in the funding tomaintain and upgrade the system.

“If we took the approach of using grantmoney to fund wireless, we’d probably beable to establish nearly ubiquitous coverageon campus,” said Voss, “but we would stillbe in a situation where we would be de-pendent on ‘miracle money’ (money left overin the budget) to maintain and upgrade theinfrastructure as it ages and applicationsbecome more bandwidth intensive. In gen-eral, counting on future miracles is a badstrategy for funding over the long term.”

IU’s present approach to general tech-nology funding shows how the processwould likely proceed for wireless. The basicapproach to establishing technology bud-gets involves collaboration between UITSand individual IU departments designed toflesh out the overall cost of an initiative overits projected life cycle. Once a cost has beenestablished, UITS provides departments witha matching grant such that the cost isroughly split between individual depart-ments and UITS.

Role and Policies of UITSThe broader role of UITS in wireless plan-

ning has been to guide departments in wire-less technology acquisition. As Voss noted,the most important aim of these guidelinesis to ensure that departments choose prod-ucts and solutions that adhere to IU’s pre-vailing IT infrastructure standards. “Thecriteria we establish for departments is thatthe wireless technology is robust, reliable,and fits within our overall environment,”

said Voss. “We essentially provide them withthe checklist, but they are responsible forthe procurement of the technology.” Anexample of the kinds of issues addressed bythese recommendations is the appropriate-ness of frequency hopping versus directsequencing in wireless infrastructure equip-ment, given IU’s security requirements.

While individual IU departments cansponsor wireless initiatives, wireless none-theless remains under the policy jurisdictionof UITS. Following are examples of key policystatements issued by UITS:� Wireless technology is not suitable for

all locations and applications and is cer-tainly not a strategic replacement for awired infrastructure. An exception is de-ployment in places where fixed wiring isnot an option because of building con-figuration, age, or location—that is,where installing traditional wiring is ei-ther not possible or not practical.

� UITS will manage all wireless hubs, ex-cept those that are mobile, temporary,or serially connected.

� All UITS-managed wireless hubs will beconnected via the VPN-secured system,unless a specific exception is granted bythe University Information TechnologyPolicy Office.

� UITS will make a site visit to assist de-partmental staff in determining the op-timal location of equipment. Where pos-sible, hubs installed by departments mustuse IU’s central, VPN-secured system.Departments must contact the UITS Net-work Operations Center to have theirwireless networks added to the VPN-secured system.

� If equipment installed by a departmentinterferes with the wireless networkmaintained for the university by UITS, theconfiguration of that equipment willhave to be changed to eliminate the con-flict, or it will have to be removed.

Page 9: Wireless Networking at Indiana University

EDUCAUSE CENTER FOR APPLIED RESEARCH 5

Wireless Networking in Higher Education Case Study 5, 2002

Wireless DeploymentIssues

Evolving standards, selection of areas fordeployment, and the extent and type ofwireless use were among the key issues re-quiring consideration.

Technology and StandardsSelection

When IU began piloting wireless in 2000,the issue of wireless standards was not ofimmediate importance, given that the IEEE802.11b standard represented the only ma-jor option at the time. However, with theimpending availability of 802.11a and802.11g, the issue of wireless standards hasgained prominence. The university is almostcertainly moving in the direction of 802.11a,with the key driver being improved band-width. However, as Voss explained, its gen-eral “newness” acts as a gating factor. “Wewill deploy 802.11a when we feel it hasmatured enough to be reliable, and the costsare such that we can reasonably replace theequipment on a standard three- or four-yearlife cycle.”

One factor that may accelerate 802.11adeployment is an interest within certain de-partments in moving more quickly towardits adoption. To ensure standardization, UITSacknowledges that it may be compelled toidentify an 802.11a-compliant product ear-lier than it had originally planned.

Scope and Focus of WirelessDeployment

In the first phase of deployment, IU de-ployed wireless in “islands” across its cam-puses, largely because of its goal ofproviding the most coverage with the leastamount of funding. The locations of theseislands—a mixture of academic classrooms,meeting places, and student gathering ar-eas—were influenced by individual depart-ments, which also provided funding. On IU’s

Bloomington campus, wireless is deployed inthe following locations:� Kelley School of Business� Carmichael Center� Creative Arts Building� Law School� Lilly Library� Indiana Memorial Union� Main Library� Spea Library� UITS

The Kelley School of Business was one ofthe earliest adopters of wireless on theBloomington campus. One key to the successof this deployment was the mandated own-ership of laptops for graduate business stu-dents, which encouraged adoption byproviding a proportionately larger pool of po-tential wireless laptop users. Another was thehigh level of importance graduate businessstudents attached to mobility—specifically theability to move from room to room withouthaving to unplug. The fact that many (if notmost) graduate business students choose touse their wireless connections even whenample wired connections are available signi-fies how embedded wireless has become intheir everyday collaboration.

Deployment of wireless in common areaswas an especially important part of IU’s ini-tial roll-out plan. The two most importantsuch areas were the Indiana Memorial Union(where wireless is now fully deployed) andthe main library. The goal of the librarydeployment was to provide greater mobilityfor students and researchers.

Primary ApplicationsFor the university as a whole, Web brows-

ing and messaging represent the dominantwireless applications. While there are relativelyfew examples of departments using wirelessfor highly specific applications, the ComputerScience Department has been relatively aggres-sive in using wireless for collaboration and ex-

Page 10: Wireless Networking at Indiana University

6

Wireless Networking in Higher Education Case Study 5, 2002

panding access to its server resources. Simi-larly, the UITS organization uses wireless in thecourse of everyday departmental activities,explained Voss, a frequent user of wireless him-self. “Wireless has been an indispensable toolin department meetings because it allows ourstaff to access and alter reports in real time,”he said. “This increases the overall productiv-ity of the department.”

Architecture andSecurity Profile

IU’s wireless implementation employs amix of Lucent 500 and 1,000 and CiscoAironet wireless access points. The locationof each access point was determined by aseries of site studies conducted by UITS. Dur-ing its site surveys, UITS determined the lo-cation and number of access points requiredto provide a given area with a throughputof at least 2 Mbps. For sites with a height-ened probability of airwave congestion, UITSused dual-transmitter access points todouble the number of simultaneous users.IU’s wireless architecture places all users ona single subnet, with a net mask that sup-ports 1,024 users. When simultaneous us-age rates begin to exceed network capacity(not expected until mid-2003), UITS expectsto split its wireless network into two domainsof 1,024 users each.

IU’s wireless network is secured througha VPN that provides both end-user authen-tication (via a user name and password) andencryption. The UITS organization managesand maintains the VPN environment, whichis designed to allow departments comingon line with wireless to simply “plug in.”The VPN solution has thus far delivered ac-ceptable security and prevented significantsecurity problems.

IU’s focus on security reflects theinstitution’s relatively high IT profile within thehigher education community, a status thathas made it a frequent target of network in-

truders. While UITS controls all aspects of thewired network (including security), the free-dom of departments to add wireless infra-structure establishes a potential point ofvulnerability in the area of network security.UITS addressed this by establishing a set ofpolicies and standards, an important elementof which is the requirement that departmentsplug into the VPN.

Perceived Benefits ofWireless

Brian Voss sees the primary benefits ofa wireless infrastructure as wider and moreconvenient access to IU’s network re-sources, beyond the range of wired bound-aries. “IU is already a top ‘most-wired’institution, and we’d like to become a top‘most-wireless’ as well,” he said. “We seethe complementary combination of wiredand wireless infrastructures as benefitingall constituencies at IU.”

While Voss sees such metrics as networkusage as a gauge of success, he believes thatthe real value of wireless is seen in morenebulous—and far-reaching—measures.“UITS looks at the value of technology interms of its potential for facilitating theteaching, learning, research, and servicemissions of the university. It’s less a ques-tion of metrics than of providing a fertileenvironment via technology.”

Lessons LearnedAlthough still in the early stages of its

wireless deployment, IU has learned severalkey lessons, the most important of whichconcerning how wireless should be intro-duced. Brad Wheeler, associate dean inUITS’s Teaching and Learning InformationTechnologies Division, sees the value of rapiddeployment (over an incremental approach)as an important lesson of IU’s experience.“With technology initiatives like wireless, it’sbetter to embrace it on a global scale,” said

Page 11: Wireless Networking at Indiana University

EDUCAUSE CENTER FOR APPLIED RESEARCH 7

Wireless Networking in Higher Education Case Study 5, 2002

Wheeler, “changing not just the technologyinfrastructure but the processes and fund-ing priorities as well.”

Brian Voss concurred and pointed to theneed to match expectations with capabili-ties—as well as with the funding requiredto back them up. “Because Indiana is aleader in IT, there was an expectation thatwe were going to build a wireless infrastruc-ture that would enable people to accessfrom anywhere on campus,” noted Voss.“This shows we needed to better set expec-tations or define an adequate fundingstream to enable ubiquitous deployment.”

IU’s wireless experience also uncoveredand debunked widely held misconceptionsabout the respective roles of the wireless andwired infrastructure, and the funding impli-cations that sprang from these misconcep-tions. The first of these was the simplisticassumption that the wireless infrastructurewould ultimately replace large componentsof the wired infrastructure, enabling largeportions of funding for the wired network tobe channeled to wireless initiatives. “Weshould have understood from the outset thatwe were talking about an augmentation de-ployment, not a substitution deployment,”explained Voss. “It would have given us amore realistic funding plan earlier on.” Thesecond (and a related) assumption that wasproved wrong was that reductions in the costof maintaining the wired infrastructure wouldfree up funds from the existing budget tosupport a broader wireless deployment.

As IU expands its embrace of wireless,one of the challenges it faces is to adaptclassroom testing practices—which oftenpermit laptop use—to a wireless comput-ing environment. The university’s short-termsolution has been to require students to re-move their wireless cards during a test. How-ever, with more and more laptops shippingwith wireless built into the motherboard, IU’sWheeler sees the challenge escalating.

“While we’re still trying to work it out, wesee two alternatives—technical solutionsand behavioral solutions. I see technical so-lutions as inferior because it turns into a ‘ra-dar and radar detector’ scenario. At the endof the day, behavioral solutions are the onlyanswer.”

The Future of Wirelessat Indiana University

Going forward, IU expects its wirelesscoverage to grow steadily and quickly. IU stillplans to build increasingly larger islands ofconnectivity, which will eventually delivercampus-wide coverage. But, in step withBrad Wheeler’s comments, the intention isto do that over a fairly short period of time.The effort began in the summer of 2002 andwill proceed rapidly in stages of deployment,culminating in a campus-wide solution forIU’s two core campuses, in Bloomington andIndianapolis, by fall 2003.

The key, said Voss, has been to find afunding model that can make broad-basedwireless deployment a reality. “We started tothink out of the box [on wireless funding],”he said. “We believe, now, we can reallocateboth one-time and ongoing funds from otherservices and savings exercises—particularlylow-speed campus network access via mo-dems, which is rapidly giving way to high-speed, user-acquired access in the form ofcable modems and DSL services.” This tran-sition means that IU will be trading its invest-ment in low-speed modem access forhigh-speed wireless access on the campuses.The shift corresponds with users’ migration,brought about by their own demands fordeveloping their home connections.

How will IU’s continued deployment becarried out? Voss said UITS will quickly evalu-ate a number of alternatives. “We have anopportunity to do the best thing we can.[That may be] a continuation of our self-deployment/management strategy, such as

Page 12: Wireless Networking at Indiana University

8

Wireless Networking in Higher Education Case Study 5, 2002

we use on our wired network, or going withan outsourced installation of access pointsand integration of that environment withour current network and security environ-ment.” He added, “Or perhaps somethingin between. Whatever it turns out to be,we’ll be determining that very quickly overthe next 60–90 days.”

On the technology front, IU continuesto evaluate IEEE 802.11a, 802.11g, and evenG3. But in the near term, as it moves to de-ploy wireless rapidly across its campuses,UITS believes it will follow the option of de-ploying 802.11b immediately. “802.11b cer-tainly works, is well established, and is

cost-effective, though it allows only a lim-ited amount of bandwidth,” Voss explained.“With a limited number of machines—which describes the current situation—802.11b will be fine. But when more andmore people are looking for wireless con-nectivity down the road, or when the nextgeneration of high-bandwidth applicationscomes along, we’ll be straining thatnetwork’s ability to deliver adequate perfor-mance. Fortunately, it looks like the newerequipment will have fairly straightforwardupgrade paths, at least within the 802.11xtechnology. So we believe we’ll be fine.”