IntroductionThe test environmentVoice over IPClient devicesTrace collectionSyslogSNMPEthernet sniffersVoIP CDR dataDefinitionsDefining mobilityChangesClientsTrafficSpecific applicationsVoIPPeer-to-peer applicationsStreaming mediaMobilityRelated workConclusions and RecommendationsFuture workREFERENCESThe Changing Usage ofa Mature Campus-wide Wireless NetworkTristan HendersonDept. of Computer ScienceDartmouth CollegeHanover, NH 03755, [email protected] KotzDept. of Computer ScienceDartmouth CollegeHanover, NH 03755, [email protected] AbyzovDept. of Computer ScienceDartmouth CollegeHanover, NH 03755, [email protected] Local Area Networks (WLANs) are now commonplaceon many academic and corporate campuses. As “Wi-Fi” technol-ogy becomes ubiquitous, it is increasingly important to understandtrends in the usage of these networks.This paper analyzes an extensive network trace from a mature802.11 WLAN, including more than 550 access points and 7000users over seventeen weeks. We employ several measurement tech-niques, including syslogs, telephone records, SNMP polling andtcpdump packet sniffing. This is the largest WLAN study to date,and the first to look at a large, mature WLAN and consider geo-graphic mobility. We compare this trace to a trace taken after thenetwork’s initial deployment two years ago.We found that the applications used on the WLAN changed dra-matically. Initial WLAN usage was dominated by Web traffic; ournew trace shows significant increases in peer-to-peer, streamingmultimedia, and voice over IP (VoIP) traffic. On-campus trafficnow exceeds off-campus traffic, a reversal of the situation at theWLAN’s initial deployment. Our study indicates that VoIP hasbeen used little on the wireless network thus far, and most VoIPcalls are made on the wired network. Most calls last less than aminute.We saw greater heterogeneity in the types of clients used, withmore embedded wireless devices such as PDAs and mobile VoIPclients. We define a new metric for mobility, the “session diameter.”We use this metric to show that embedded devices have differentmobility characteristics than laptops, and travel further and roam tomore access points. Overall, users were surprisingly non-mobile,with half remaining close to home about 98% of the time.Categories and Subject DescriptorsC.2.1 [Network Architecture and Design]: Wireless communica-tion; C.2.3 [Network Operations]: Network monitoringGeneral TermsMeasurementPermission to make digital or hard copies of all or part of this work forpersonal or classroom use is granted without fee provided that copies arenot made or distributed for profit or commercial advantage and that copiesbear this notice and the full citation on the first page. To copy otherwise, torepublish, to post on servers or to redistribute to lists, requires prior specificpermission and/or a fee.MobiCom’04, Sept. 26-Oct. 1, 2004, Philadelphia, Pennsylvania, USA.Copyright 2004 ACM 1-58113-868-7/04/0009 ...$5.00.KeywordsWireless network, Wi-Fi, 802.11, voice, VoIP, telephony, WLAN1. INTRODUCTIONWireless Local Area Networks (WLANs) have become com-monplace, especially on university and corporate campuses, andincreasingly in public “Wi-Fi hotspots” as well. Most modernlaptops are equipped with a network adapter that can access oneor more types of IEEE 802.11 network, but wireless devices arerapidly diversifying to include PDAs, printers, audio players, andmore. These new devices lead to changes in the way that WLANsare used. For instance, we might expect a wireless PDA to have dif-ferent usage patterns than a wireless printer; a PDA might be moremobile as its user traverses a WLAN-enabled campus, whereas theprinter may remain in one place to serve wireless clients.The growing popularity of WLANs encourages the developmentof new applications, which may also exhibit new usage charac-teristics. Real-time multimedia applications, for example, havequality-of-service (QoS) requirements that may be difficult to fulfillin a shared-medium WLAN. Some of these new applications anddevices may emerge simultaneously; for instance many wirelessPDAs are sold equipped with streaming audio or video software.Understanding the usage, and trends in usage, of these new de-vices and applications is important for providers who deploy andmanage WLANs, for designers who develop new high-throughputand multimedia-friendly wireless networking standards, and forsoftware developers who create new wireless and location-awareapplications.In this paper we study a large trace of network activity in a ma-ture production wireless LAN. Dartmouth College has had 802.11bcoverage for three years in and around nearly every building oncampus, including all administrative, academic, and residentialbuildings, as well as most social and athletic facilities. We collectedextensive trace information from the entire network throughout theFall and Winter terms of 2003/2004.Our work expands significantly upon previous studies. Tangand Baker [19] traced 74 computer-science clients in one build-ing for 12 weeks. There are two more recent studies; Schwab andBunt [18] examine 134 users over one week, and Chinchilla et al.track over 7,000 wireless cards for 11 weeks at UNC, examiningweb-browsing activity (for 4 weeks) and location prediction. Ourearlier study, conducted at Dartmouth in 2001 [11], looked at morethan 1700 users over 11 weeks. In this new 2003/04 study, weobserved over 7,000 unique wireless cards using over 550 accesspoints over the course of a 17-week trace period.In particular, our study extends previous work by examiningtrends in behavior of a mature WLAN, and by examining geo-graphic mobility within a large WLAN. We compare this 2003/4trace to our earlier trace from Fall 2001, taken shortly after the ini-tial installation of our campus WLAN. We found that the workloadhas changed significantly since 2001, and is significantly differentthan in other previous studies. We saw new embedded wirelessdevices, and new applications such as peer-to-peer services andstreaming multimedia.We next describe the environment of our study, the DartmouthCollege campus, and then detail our tracing methodology in Sec-tion 3. In Section 4 we present and compare the most interestingcharacteristics of the data to those taken from an earlier study dur-ing the initial WLAN deployment. In Section 5 we examine threeparticular applications in detail: peer-to-peer file sharing, stream-ing