Monitoring 3G cellular networks

Here follows a (still growing) list of papers dealing with measurements in 3G cellular networks.

Passive monitoring of TCP performances

  • The paper [benko-infocom-04] was the first to report large-scale measurements of TCP RTT from a real cellular networks (GPRS) based on handshake packets.
  • [vacirca-wicon-05] reported large-scale measurements of TCP RTT and RTO for an operational UMTS network, based on the analysis of DATA/ACK pairs (not only handshake packets).
  • more recently [romirer-tma-09] reported network-wide RTT measurements from GPRS/EDGE and UMTS/HSPA network, with data from Q1-2009.

Active measurements

  • [pentikousis-letter-05] shows throughput measurements from a public UMTS network.
  • [kilpi-networking-06] analyses the temporal variability of RTT.
  • [fabini-networking-09] looks at patterns of HSPA delays across three different networks.
  • [barbuzzi-letter-08] leverages active measurements to infer the setting of some internal network parameters (e.g. timeouts) and channel assignment/release delays.
  • [barbuzzi-tma2010desrto] introduce an algorithm and a tool able to identify SRTOs in TCP connections.




  • [benko-infocom-04] bibtex
    P. Benko, G. Malicsko, and A. Veres, "A large-scale, passive analysis of end-to-end TCP performance over GPRS." 2004, pp. 1882-1892 vol.3.
    @inproceedings{benko-infocom-04, abstract = {In this paper a passive methodology for TCP performance evaluation over general packet radio service (GPRS) networks is presented that relies on traffic monitoring at the GPRS ingress/egress router interface (Gi). Based on the IP and TCP headers of the packets we estimate the end-to-end performance of TCP connections such as connection setup behavior and data transfer goodput. In order to identify the effects behind the measured performance the introduced algorithms estimate round trip delays, packet loss ratios, available channel rates, throughput and cany out bottleneck analysis. Large-scale GPRS measurements in seven countries are presented to analyze TCP performance and demonstrate the applicability of the method. The effects of different TCP parameters such as maximum segment size, selective acknowledgements, timestamp usage and receiver window size are also quantified. GPRS measurement results are compared to a wireline dial-up network to identify the effects specific to the wireless environment},
      author = {Benko, P. and Malicsko, G. and Veres, A.},
      doi = {10.1109/INFCOM.2004.1354598},
      interhash = {e011ccbe03c7dee6b36e554603238387},
      intrahash = {fe11244effebf2819935bfa14ccd4129},
      issn = {0743-166X},
      journal = {INFOCOM 2004. Twenty-third AnnualJoint Conference of the IEEE Computer and Communications Societies},
      month = {March},
      pages = {1882-1892 vol.3},
      title = {A large-scale, passive analysis of end-to-end TCP performance over GPRS},
      volume = 3, year = 2004 }
  • [vacirca-wicon-05] bibtex
    F. Vacirca, F. Ricciato, and R. Pilz, "Large-scale RTT measurements from an operational UMTS/GPRS network." 2005, pp. 190-197.
    @inproceedings{vacirca-wicon-05, abstract = { In this paper we present some observations about TCP RTT as captured in the live traffic of an operational GPRS/UMTS network. RTT samples are extracted from traces collected by passive monitoring the Gn interfaces of one of the major providers in Austria, EU. We compare results for GPRS and UMTS, and expose some methodological issues involved in performing this type of analysis, for instance the potential bias introduced by heavy users. We also explore the correlation of RTT with time-of-day. The primary motivation for this research is to gain a better understanding of the dominant causes that shape the behaviour and the distribution of RTT across the TCP connections that populate a GPRS/UMTS network. The results presented here are useful for the scientific community, for instance to set more realistic simulation scenarios for other research activities. In addition, RTT measurements as obtained by large-scale passive monitoring can be used in combination with other indicators to build a summary indicator of the performances experienced by the TCP users at the scale of the whole network, i.e. a global RTT-based KPI (key performance indicator).},
      author = {Vacirca, F. and Ricciato, F. and Pilz, R.},
      doi = {10.1109/WICON.2005.19},
      interhash = {778707d95a6ba5e69acdb2af60a3e596},
      intrahash = {2f69363df6c8ac54cbcc83f53a73c100},
      journal = {Wireless Internet, 2005. Proceedings. First International Conference on},
      month = {July},
      pages = { 190-197},
      title = {Large-scale RTT measurements from an operational UMTS/GPRS network},
      year = 2005 }
  • [romirer-tma-09] bibtex
    P. Romirer-Maierhofer, F. Ricciato, A. D’Alconzo, R. Franzan, and W. Karner, "Network-Wide Measurements of TCP RTT in 3G," Traffic Monitoring and Analysis, pp. 17-25, 2009.
    @article{romirer-tma-09, abstract = {In this study we present network-wide measurements of Round-Trip-Time (RTT) from an operational 3G network, separately for GPRS/EDGE and UMTS/HSxPA sections. The RTTs values are estimated from passive monitoring based on the timestamps of TCP handshakingpackets. Compared to a previous study in 2004, the measured RTT values have decreased considerably. We show that the network-wideRTT percentiles in UMTS/HSxPA are very stable in time and largely independent from the network load. Additionally, we presentseparate RTT statistics for handsets and laptops, finding that they are very similar in UMTS/HSxPA. During the study we identifieda problem with the RTT measurement methodology — mostly affecting GPRS/EDGE data — due to early retransmission of SYNACK packetsby some popular servers.},
      author = {Romirer-Maierhofer, Peter and Ricciato, Fabio and D’Alconzo, Alessandro and Franzan, Robert and Karner, Wolfgang},
      interhash = {0e42abf2fb49b376e608e7dc54f371bc},
      intrahash = {c96dd04723d1180b27fa3417db414528},
      journal = {Traffic Monitoring and Analysis},
      pages = {17--25},
      title = {Network-Wide Measurements of TCP RTT in 3G},
      url = {},
      year = 2009 }
  • [kilpi-networking-06] bibtex
    J. Kilpi and P. Lassila, "Micro- and Macroscopic Analysis of RTT Variability in GPRS and UMTS Networks," Networking 2006. Networking Technologies, Services, and Protocols; Performance of Computer and Communication Networks; Mobile and Wireless Communications Systems, pp. 1176-1181, 2006.
    @article{kilpi-networking-06, abstract = {We study the data from a passive TCP/IP traffic measurement from a Finnish operator’s GPRS/UMTS network. Of specific interest is the variability of Round Trip Times (RTTs) of TCP flows. The RTTs are analysed at micro- and macroscopic level. The microscopic level involves detailed analysis of the RTTs of individual flows, and we are able to detect, e.g., periodic behavior (via Lomb periodogram) and rate changes in the radio channel. At the macroscopic level we focus on the impact of so called self-congestion caused by bandwidth sharing at the mobile device itself, and it is shown how this seriously affects the RTTs observed by a given flow, both in GPRS and in UMTS. ER -},
      author = {Kilpi, Jorma and Lassila, Pasi},
      interhash = {2d2595c981b57b3bd616b3efca5e1423},
      intrahash = {2a6cae24fd2d65b3b88b902c5d834555},
      journal = {Networking 2006. Networking Technologies, Services, and Protocols; Performance of Computer and Communication Networks; Mobile and Wireless Communications Systems},
      pages = {1176--1181},
      title = {Micro- and Macroscopic Analysis of RTT Variability in GPRS and UMTS Networks},
      url = {},
      year = 2006 }
  • [fabini-networking-09] bibtex
    J. Fabini, W. Karner, L. Wallentin, and T. Baumgartner, "The Illusion of Being Deterministic – Application-Level Considerations on Delay in 3G HSPA Networks," NETWORKING 2009, pp. 301-312, 2009.
    @article{fabini-networking-09, abstract = {The delay experienced by mobile applications in HSPA networks depends to a large extent on highly dynamical global context like, e.g., cell load or algorithms and thresholds governing radio resource scheduling, and on local context like, e.g., user-generatedload or load history. These complex uncertainty factors are outside of an application’s sphere of influence and result inapplications perceiving HSPA link behavior as non-deterministic and non-reproducible.},
      author = {Fabini, Joachim and Karner, Wolfgang and Wallentin, Lukas and Baumgartner, Thomas},
      interhash = {03cb828b834597fa3ec3dacf0f53a033},
      intrahash = {34c8a0936a5a1902925f87b86dcf2751},
      journal = {NETWORKING 2009},
      note = {work in progress},
      pages = {301--312},
      title = {The Illusion of Being Deterministic – Application-Level Considerations on Delay in 3G HSPA Networks},
      url = {},
      year = 2009 }
  • [barbuzzi-letter-08] bibtex
    A. Barbuzzi, F. Ricciato, and G. Boggia, "Discovering Parameter Setting in 3G Networks via Active Measurements," Communications Letters, IEEE, vol. 12, iss. 10, pp. 730-732, 2008.
    @article{barbuzzi-letter-08, abstract = {The behavior and performance of a UMTS network are governed by a number of parameter settings that are configured by the network operator, e.g., timeouts. In this letter we show that the actual value of such parameters can be inferred by a conceptually simple set of end-to-end measurements, without any cooperation with the network operator. In principle, such information can be used by researchers to define realistic network scenarios, e.g., for their simulations. Moreover, it can be used by a malicious attacker to fine-tune a large scale attack against the radio access network, e.g., a paging attack.},
      author = {Barbuzzi, A. and Ricciato, F. and Boggia, G.},
      doi = {10.1109/LCOMM.2008.080913},
      interhash = {fd4b48bd63cdc7c1143ba098914bee66},
      intrahash = {a1f2683dff752c5073d090f88267a3dd},
      issn = {1089-7798},
      journal = {Communications Letters, IEEE},
      month = {October },
      number = 10, pages = {730-732},
      title = {Discovering Parameter Setting in 3G Networks via Active Measurements},
      volume = 12, year = 2008 }
  • [pentikousis-letter-05] bibtex
    K. Pentikousis, M. Palola, M. Jurvansuu, and P. Perala, "Active goodput measurements from a public 3G/UMTS network," Communications Letters, IEEE, vol. 9, iss. 9, pp. 802-804, 2005.
    @article{pentikousis-letter-05, abstract = { Although third generation cellular networks are being deployed in many countries, rigorous measurement studies from public networks have yet to be published. After extensive experimentation, we collect measurements from a commercial WCDMA 3G/UMTS network and observe that the goodput of the first of a series of back-to-back transfers is consistently below par. We compare these results with those obtained from an all-wired network and show that 3G signaling plays an important, albeit detrimental role in end-user goodput. Finally, we illustrate that near-nominal goodput is realizable for large payloads only.},
      author = {Pentikousis, K. and Palola, M. and Jurvansuu, M. and Perala, P.},
      doi = {10.1109/LCOMM.2005.1506708},
      interhash = {25a5ebb5bc182aa2e6ee3f2e79ce11a5},
      intrahash = {9b5de5881cee616df675a83c74d2854e},
      issn = {1089-7798},
      journal = {Communications Letters, IEEE},
      month = Sep, number = 9, pages = { 802-804},
      title = {Active goodput measurements from a public 3G/UMTS network},
      volume = 9, year = 2005 }
  • [barbuzzi-tma2010desrto] bibtex
    A. Barbuzzi, G. Boggia, and L. Grieco, "DeSRTO: An Effective Algorithm for SRTO Detection in TCP Connections," Traffic Monitoring and Analysis, pp. 87-100, 2010.
    @article{barbuzzi-tma2010desrto, abstract = {Spurious Retransmission Timeouts in TCP connections have been extensively studied in the scientific literature, particularlyfor their relevance in cellular mobile networks. At the present, while several algorithms have been conceived to identifythem during the lifetime of a TCP connection (e.g., Forward-RTO or Eifel), there is not any tool able to accomplish the taskwith high accuracy by processing off-line traces. The only off-line existing tool is designed to analyze a great amount oftraces taken from a single point of observation. In order to achieve a higher accuracy, this paper proposes a new algorithmand a tool able to identify Spurious Retransmission Timeouts in a TCP connection, using the dumps of each peer of the connection.The main strengths of the approach are the great accuracy and the absence of assumptions on the characteristics of TCP protocol.In fact, except for rare cases that are not classifiable with absolute certainty at all, the algorithm shows no ambiguousnor erroneous detections. Moreover, the tool is also able to deal with reordering, small windows, and other cases where competitorsfail. Our aim is to provide to the community a very reliable tool to: (i) test the working behavior of cellular wireless networks,which are more prone to Spurious Retransmission Timeouts with respect to other technologies; (ii) validate run-time SpuriousRetransmission Timeout detection algorithms.},
      author = {Barbuzzi, Antonio and Boggia, Gennaro and Grieco, Luigi},
      interhash = {48d72a138b1bd98bc75ba2e3f505a560},
      intrahash = {72bab3e0486372e6191d2c65e2dcf4c6},
      journal = {Traffic Monitoring and Analysis},
      pages = {87--100},
      title = {DeSRTO: An Effective Algorithm for SRTO Detection in TCP Connections},
      url = {},
      year = 2010 }