A. Helmy's picture NOMADS group logo

Ahmed Helmy
Professor & Graduate Coordinator
Computer & Information Science & Engineering (CISE) Department
Founder and Director: Mobile Networking Laboratory (NOMADS group)
College of Engineering, University of Florida
Gainesville, FL 32611
email: helmy at ufl.edu

MobiLib: Community-wide Library of Mobility and Wireless Data
New: MobiLib2.0 is in the works through this link, including the Flutes vs Cellos study.

UFL Wireless Coverage Map - Note: If you use any of these tools, code or data, please make sure to cite our correspnding paper(s)! Thanks!
- Note2: Please abide by all the licenses under which this data and code are released.



This webpage aims to establish a community-wide library of mobile wireless networks traces and measurements. The goal is to have the traces, simulation code, test-suites (e.g., benchmarks, test scenarios) and models (of mobility, traffic and user behavior) established by the experts in the field, widely available for everyone to use and compare against.

History: Initial idea for MobiLib started in 2003 by A. Helmy and his group at USC. Actual effort for initial data collection and library setup started in 2004 by A. Helmy and Fan Bai. First launch of the MobiLib website was in 2005 by A. Helmy, F. Bai and Wei-Jen Hsu. Many Prof.s and researchers around the world encouraged this work and provided links to datasets (acknowledged at the bottom of this page). Around the same time frame Prof. D. Kotz (Dartmouth) was starting his effort for CRAWDAD (the goal for which was to 'house' the data in addition to collection from Dartmouth, unlike MobiLib which was to provide pointers to existing datasets). Most wireless datasets contributed by various Prof.s, including our group, are housed (wholly or partially) in CRAWDAD. MobiLib moved to the University of Florida in '06/'07, where A. Helmy and his group are conducting further research in the area of mobility and behavioral modeling for wireless network users.

Thanks to many major universities who agreed to provide traces (or pointers to traces) including USC, UFL, MIT, UCSD, Dartmouth, UCSB, UIUC, GA Tech, Purdue, UCLA, Rice, Boston U, Columbia, U Washington, UNC. Links are continuously updated. If you are interested and can contribute, please contact: helmy@ufl.edu.

Download Traces
(Getting access to measurements)
Related Work by NOMADS Group
(Projects, Publications)
Related Projects
Related Publications

Pointers to traces

  1. (Apr. 18) New: WLAN, netflow and vehicular mobility traces to be added in the future at this link.
  2. (Dec. 17) Note: The "Nile" server @UF is down/out-of-date and needs updating... to be updated soon, stay tuned! Thanks!
  3. (July 08) Parts of the USC trace have been contributed to the Crawdad project @Dartmouth.
  4. (Aug. 12, 05) USC WLAN trace (VPN sessions, DHCP, Traps, Flow size) ([Sep 20]: Longer processed trace up to Aug '05 available here)
  5. (March 28, 06) USC traces based on surveys and observations
  6. CRAWDAD: Wireless LAN traces of Dartmouth College (syslog, SNMP, and tcpdump data) [CRAWDAD has grown over the years to include many other traces including ours from USC].
  7. UCSD: PDA trace of University of California, San Diego (contains access points seen by PDAs taken at a 20 seconds interval, for 3 months) [check CRAWDAD for availability].
  8. Wireless LAN trace of MIT (SNMP data from 3 engineering buildings for a month)
  9. UCSD: Wireless LAN trace of Sigcomm, 2001 [check CRAWDAD for availability].
  10. Bus schedule traces (.tar.gz) from the Houston Metropolitan area from the Monarch project at Rice University
  11. Encounter traces from the Haggle project (traces of Bluetooth sightings by groups of users carrying small devices (iMotes)).
  12. Wireless traces at SIGCOMM 04 from U. Washington (Used in this paper)
  13. Mobile phone traces from MIT Media Lab (and the Reality Mining project)
  14. Traces from the roof-top network at MIT
  15. (Nov 27, 05) IETF traces from UCSB (Used in this paper)
  16. (Jan 31, 05) Bus (DTN) traces from UMass Amherst (Used in this paper)
  17. (March 28, 06) Wireless LAN traces from UNC (University of North Carolina) and its surroundings. [Check CRAWDAD]
  18. (March 29, 06) Vehicular network traces from Georgia Tech, Atlanta. [Check CRAWDAD]

Related Work and Publications by the NOMADS group

- Here are selected publications from the NOMADS group at UF (and previously at USC): [Note: For more recent publications check A. Helmy's home page and related publications links.]

  1. - New: (Nov '11) SOM: "Multidimensional Modeling and Analysis of Wireless Users Online Activity and Mobility: A Neural-networks Map Approach" (Self Organizing Maps), ACM MSWiM, Nov. '11 (and IEEE INFOCOM '11).
  2. - New: (Sept '11) The TRACE framework for Mining Mobile User Behavior: - IEEE Transactions on Mobile Computing (TMC) journal version [accepted Sept. '11].
  3. - New: (Aug '11) CSI: Profile-cast, a new communication paradigm for mobile social networks: - Ad Hoc Networks - Elsevier Journal [accepted Aug. '11], - Watch the videos.
  4. - The IMPACT study on Nodal Encounter Patterns in WLAN Traces: - IEEE Transactions on Mobile Computing (TMC) [published Nov. '10].
  5. - Mobile web usage co-clustering: "Data-driven Co-clustering Model of Internet Usage in Large Mobile Societies", ACM MSWiM, Oct 2010.
  6. - Time-variant Community (TVC) mobility model: - IEEE/ACM Transactions on Networking (ToN) [published Oct. '09] (and IEEE INFOCOM '07) , - mobility tool
  7. - The IMPORTANT mobility study: - IEEE INFOCOM '03 , - Ad Hoc Network Journal Nov. '03 , - mobility tool

- Further related work from the NOMADS group: (most papers available through this publications website)

  1. - U. Kumar, N. Yadav, and A. Helmy, "Analyzing Gender-gaps in Mobile Student Societies," CRAWDAD Workshop poster (colocated with MOBICOM 2007)
  2. - J. Kim, Y. Du, M. Chen, and A. Helmy, "Comparing Mobility and Predictability of VoIP and WLAN Traces," CRAWDAD Workshop poster (colocated with MOBICOM 2007)
  3. - U. Kumar, N. Yadav, and A. Helmy, "Gender-based feature analysis in Campus-wide WLANS," MOBICOM 2007 poster and SRC. [A webpage for the results is available HERE]
  4. - W. Hsu, D. Dutta, and A. Helmy, "Profile-cast: Behavior-Aware Mobile Networking," MOBICOM 2007 poster and SRC.
  5. - W. Hsu, D. Dutta, and A. Helmy, "Extended abstract: Mining behavioral groups in large wireless LANs," to appear in Proceedings of MOBICOM 2007. [Longer version of technical report available HERE] Highlights: In this work we leverage unsupervised learning technique (i.e., clustering) to identify groups of users with distinct behavioral patterns in the WLAN traces. We develop the TRACE framework and use summarized mobility pattern as an example to show the applicability of the framework. We find that university campus is a diverse setting in which hundreds of groups with distinct behavioral modes exist, and the group sizes follow a power-law distribution.
    The technique we propose in this paper could be used for better mobility models, behavioral norm establishment and abnormality detection, profile-based services such as advertisement and group-cast, to name a few.

  6. - W. Hsu, T. Spyropoulos, K. Psounis, and A. Helmy, "Modeling Time-variant User Mobility in Wireless Mobile Networks," in Proceedings of IEEE INFOCOM, May 2007. [Webpage for the time-variant community model HERE, mobility trace generator and manual available] Highlights: The Time-Variant Community Mobiliy Model is a model we create to capture two important mobility characteristics we observed earlier from WLAN traces. These two mobility characteristics are skewed location visiting preferences and periodical re-appearance.
    While improving the realism of the mobility model, we also keep mathematical tractability as a requirement for the mobility model. We use random-direction mobility model as the basic building block, modify it to incorporate fore-mentioned mobility chracteristics. We are able to derive two quantities of interest related to mobility-assisted routing, the hitting time and the meeting time. We intereted in deriving other quantities in the future.
    We make the code for our time-variant community model available here. The code has many parameters and provides full flexibility to match with various mobility scenarios (for full details, refer to the manual). It simulates the hitting time, the meeting time, and prints the movement traces in two option formats: (1) NS-2 compatible format, or (2) time, location (in x,y coordinates) format.

  7. - W. Hsu, D. Dutta, and A. Helmy, "Mobicom Poster Abstract: On the Structure of User Association Patterns in Wireless LANs," to appear in Mobile Computing and Communication Review. Earlier version of poster abstract accepted by MOBICOM 2006.
  8. - W. Hsu, A. Helmy, "On Nodal Encounter Patterns in Wireless LAN Traces", The 2nd IEEE Int.l Workshop on Wireless Network Measurement (WiNMee), April 2006.
  9. - W. Hsu, A. Helmy, "On Modeling User Associations in Wireless LAN Traces on University Campuses", The 2nd IEEE Int.l Workshop on Wireless Network Measurement (WiNMee), April 2006.
  10. - W. Hsu, A. Helmy, "IMPACT: Investigation of Mobile-user Patterns Across University Campuses using WLAN Trace Analysis", USC Technical Report (under submission), July 2005. [Longer technical report (update July 18, 05). IMPACT webpage]. [This paper includes most information in the 2 Winmee papers above. To be extended further in the future.]

    Highlights: This paper provides the most comprehensive study of WLAN traces to date. Traces collected from four major universities (~12,000 users) are analyzed using metrics for individual user and group behaviors. Similarities and differences across campuses are studied. Conclusions provide great insight into realistic behavior of wireless users. Most users are 'on' for a small fraction of the time, number of access points visited (per user) is quite low, and on-line user mobility is quite low. On average, a user encounters only 2%-6% of the user population. Encounter-graphs and small worlds are introduced to model encounter patterns between users. We find that number of encounters follows a biPareto distribution and the frienship indexes follow exponential distributions. A paradigm for 'encounter-based information diffusion' is introduced for efficient data dissemination in mobile networks.

  11. - W. Hsu, A. Helmy, "Principal Component Analysis of User Association Patterns in Wireless LAN Traces", IEEE INFOCOM poster, April 2006.
  12. - W. Hsu, A. Helmy, "Capturing User Friendship in WLAN Traces", IEEE INFOCOM poster, April 2006.
  13. - W. Hsu, A. Helmy, "Analyzing Principal Characteristics of User Association Patterns and Eigen-behavior in Wireless LAN Traces", November 2005. [Submitted for review]
  14. - F. Bai, A. Helmy, "Impact of Mobility on Mobility-Assisted Information Diffusion (MAID) Protocols", IEEE SECON 2007.
    Highlights: This study analyzes a class of protocols, MAID, that utilize mobility for information diffusion. MAID uses encounter information to create age gradients towards the target, and can be used for discovering resources, routing or locating nodes efficiently in future mobile networks. Analytical models are developed to evaluate MAID's performance during its various (transient and steady-state) phases of operation. Extensive simulations are used to validate these models and to study the sensitivity of MAID to a rich set of mobility models. We find that although MAID is sensitive to the mobility pattern, its steady state performance is, surprisingly, insensitive to velocity. We identify the properties of the 'age gradient tree' as the key factor to explain this interplay between mobility and the MAID protocols.

  15. - Wei-jen Hsu and Ahmed Helmy, "Encounter-based Message Broadcasting in Ad Hoc Networks with Intermittent Connectivity", ACM MOBIHOC poster, May 2005. [Poster slides, Extended abstract, Initial technical report]

  16. - The IMPORTANT mobility simulation and analysis tool. (Can be directly used for NS-2 simulations)

  17. - The Weighted waypoint (WWP) mobility model.

Related projects

Related publication on user-bahavior analysis and modeling in WLANs (other, more recent publications may be added in the future)

Thanks to the people who have contributed traces and/or encouraged this work: Mostafa Ammar, Richard Fujimoto (Georgia Tech), Kevin Almeroth, Elizabeth Royer (UCSB), David Kotz, Andrew Campbell (Dartmouth), Nitin Vaidya, Jennifer Hou (UIUC), Ness Schroff, Sonia Fahmy (Purdue), Mario Gerla, Medy Sanadidi (UCLA), Tracy Camp (Colorado School of Mines), David Wetherall (U. Washington), Victor Bahl (Microsoft Research), Ed Knightly, David Johnson (Rice), Rene Cruz (UCSD), Maria Papadopouli, Kevin Jeffay (U North Carolina), Henning Schulzrinne (Columbia), Azer Bestavros, Ibrahim Matta (Boston U), Dina Katabi (MIT), Stefano Basagni (Northeastern U.), Michele Zorzi (U. Padova/UCSD), Eylem Ekici (Ohio State U), Jim Kurose, Brian Levine (U. Mass - Amherst), Srikanth Krishnamurthy, Michalis Faloutsos (UC Riverside)

This material is based upon work supported in part by the National Science Foundation under Grant No. 0134650.
Any opinions, findings and conclusions or recomendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF).

Privacy: Anonymization techniques have been used to remove information that may help identify MAC addresses of devices. No private information is contained within these traces. The trace collection process was conducted in adherance with the code of the corresponding universities, and proper permissions were granted as/when needed.

Number of visitors since Jul. 18 2005:
This page was created May '05 at USC. Moved to UF 2007. Updated intermittently, with the latest update 2011...
New, on-going updates, Dec. 2017... stay tuned!

[Apr 2013] This page has been translated by students/volunteers in the Czech Republic to make it available for researchers and scientists there.