The case for an Open and Scalable Internet Measurements Service B. Donnet, O. Bonaventure Universitcatholique de Louvain (UCL), Belgium Objective The Internet was designed with a best effort service in mind. However, nowadays the Internet is more and more used to carry non-best effort traffic and many content providers and Internet users are trying to optimise their transmission of IP packets by using various techniques such as: - overlay networks - content distribution networks - server selection We expect that with the growing importance of the internet combined with the growth of multihoming, more and more Internet users will need to better select the paths taken by their packets or the server from which they receive information. Today, although many of measurements techniques have been developped within the IPPM working group of the IETF, an application that needs to select a path or a server must implement its own measurement system to obtain data to perform its selection. This implies that several applications running on the same host or in the same campus will probably perform almost the same kind of measurements. In the long term, duplicating those measurements is not the appropriate solution. A better solution would be to develop and Open and Scalable Internet Measurements Service (OSIMS). This service would run continuously and could be queried by applications requiring measurement information. It could be composed of the following components : - a topology map of the (visible part of the) Internet. Existing tools are, for instance, skitter [3], RIPE NCC TTM [4], scamper [5] or DIMES [6] - delay information. Existing tools are Ping [15], NIMI [16] and OWAMP [17]. - geolocation of host. Existing tools are, for instance, GeoPing [13] CBG [11], GeoBud [12] and Octant [14] - bandwidth information. Existing tools are, for instance, PathLoad [7], PathRate [8], BRoute [9] and PathChar [10]. As mentionned above, there exist plenty of tools for measuring the internet. However, they are all working in isolation from each others. Even though efforts are currently done on efficient topology measurement [2], we do not already know how to collect topology, delay and bandwidth information in a scalable and secure way. For delay information, coordinates can be a valid way of distributing delay information in a scalable way, but recent work indicates that security issues can be a problem if one is willing to deploy coordinates as an always on service [1]. We believe that such a measurement service will be a key component of the future Internet and that a testbed such as onelab++ would be the perfect project to develop and validate it. References [1] M. A. Kaafar, L. Mathy, T. Turletti and W. Dabbous. "Virtual Networks under Attack: Disrupting Internet Coordinate Systems", In Proc. ACM CoNEXT, Dec. 2006. [2] B. Donnet, P. Raoult, T. Friedman and M. Crovella. "Deployment of an Algorithm for Large-Scale Topology Discovery", In IEEE Journal on Selected Areas in Communications, Sampling the Internet: Techniques and Applications, 24(12), pp. 2210 - 2220, Dec. 2006. [3] B. Huffaker, D. Plummer, D. Moore and k. claffy. "Topology Discovery by Active Probing", In Proc. Symposium on Applications and the Internet (SAINT), Jan. 2002 [4] F. Georgatos, F. Gruber, D. Karrenberg, M. Santcroos, A. Susanj, H. Uijterwaal and R. Wilhelm. "Providing Active Measurements as a Regular Service for ISPs", In Proc. Passive and Active Measurement Workshop (PAM), Apr. 2001. [5] M. Luckie. "IPv6 Scamper". WAND Network Research Group. See http://www.wand.net.nz/~mjl12/ipv6-scamper/. 2005 [6] Y. Shavitt and E. Shir. "DIMES: Let the Internet Measures Itself", In ACM SIGCOMM Computer Communication Review, 35(5), pp. 71 - 74, 2005. [7] M. Jain and C. Dovrolis. "End-to-End Available Bandwidth: Measurement Methodology, Dynamics and Relation with TCP Throughput", In IEEE/ACM Transactions in Networking, 11(4), pp. 537 - 549, Aug. 2003. [8] C. Dovrolis, P. Ramanathan and D. Moore. "Packet Dispersion Techniques and a Capacity Estimation Methodology", In IEEE/ACM Transactions on Networking, 12(6), pp. 963 - 977, Dec. 2004. [9] N. Hu and P. Steenkiste. "Exploiting Internet Route Sharing for Large-Scale Available Bandwidht Estimation", In Proc. ACM/USENIX Internet Measurement Conference (IMC), Oct. 2005. [10] V. Jacobson. "PathChar". See ftp://ftp.ee.lbl.gov/pathchar. Apr. 1997. [11] B. Gueye, A. Ziviani, M. Crovella and S. Fdida. "Constraint-Based Geolocation of Internet Hosts", In IEEE/ACM Transaction on Networking, 14(6), pp. 1219 - 1232, Dec. 2006. [12] B. Gueye, S. Uhlig, A. Ziviani and S. Fdida. "Leveraging Buffering Delay Estimation for Geolocation of Internet Hosts", In Proc IFIP Networking, May 2006. [13] V. N. Padmanabhan and L. Subramanian. "An Investigation of Geographic Mapping Techniques for Internet Hosts", In Proc. ACM SIGCOMM, Aug. 2001 [14] B. Wong, I. Stoyanov and E. Sirer. "Geolocalization on the Internet Through Constraint Satisfication", in Proc. Workshop on Real, Large Distributed Systems (WORLDS), Nov. 2006. [15] M. Muuss. "Packet Internet Groper". See Man Pages. [16] A. Adams, J. Mahdavi and M. Mathis and V. Paxson. "Creating a Scalable Architecture for Internet Measurement", In Proc. INET, Jul. 1998. [17] S. Shalunov, B. Teitelbaum, A. Karp, J. Boot and M. J. Zekauskas. "A One-Way Active Measurement Protocol (OWAMP)", Internet Engineering Task Force, RFC 4656, Sept. 2006.