On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data Alessandro Improta Advisors: Prof. Luciano Lenzini Prof. Gigliola Vaglini Ing. Enrico Gregori Pisa - May 24th, 2013 Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
The Internet role in society The Internet is becoming day by day more important for everyone Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
... but what truly is the Internet? Nevertheless, the Internet is perceived by every user as a magic box What happens to my data packets once they leave my home router? Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
First... a little bit of history 1969 - ARPANET 1985 - NSFNET 1995 - Commercial Internet ARPANET was one of the first network to implement TCP/IP (1983) Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
First... a little bit of history 1969 - ARPANET 1985 - NSFNET 1995 - Commercial Internet NSFNET use from for-profit organizations was acceptable when it was in support of open research and education Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
What about today? 1969 - ARPANET 1985 - NSFNET 1995 - Commercial Internet From then, its real structure became hidden, as well as its potential structural weaknesses Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
Why is important to discover the Internet structure? To understand how packets are routed in the Internet To understand how to optimize Internet paths by analyzing existing deficiencies To develop more scalable interdomain routing protocols and architectures To construct economy-based models of the global Internet growth To develop better topology generators to simulate the Internet To select data centers for server replicas by taking into account the Internet paths To properly select peers or upstream providers based on their connectivity Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
What researchers can do? There not exist any dedicated tool to discover the Internet topology at any level... ... but there exist two main workarounds: traceroute (IP-level) BGP route collectors (AS-level) Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
BGP Route Collectors A Route Collector (RC) is a device which collects BGP routing data from co-operating ASes. Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
The Internet AS-level topology “An AS is a connected group of one or more IP prefixes run by one or more network operators which has a single and clearly defined routing policy” . [RFC 1930] 44,389 AS numbers and 170,204 inter-AS connections were found in January 2013 topology Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
PhD Contribution 1 Inter-AS economic relationship inference from BGP data 2 Geographic AS path inference from BGP data 3 Quantification of BGP data completeness 4 Identification of ideal BGP feeders to improve current situation Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
PhD Contribution 1 Inter-AS economic relationship inference from BGP data 2 Geographic AS path inference from BGP data 3 Quantification of BGP data completeness 4 Identification of ideal BGP feeders to improve current situation Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
BGP export policies Internet AS-level topology is not an undirected graph provider-customer : the customer pays the provider to reach every AS peer-to-peer : the two ASes exploits each other to reach their customer-cones (typically free-of-charge) sibling-to-sibling : each AS acts as a provider for the other Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
AS path valley-free property In other words, ASes should not transit traffic between ✗ two of its peers ✗ two of its providers ✗ a peer and a provider A p2c edge can be followed by only p2c or s2s edges A p2p edge can be followed by only p2c or s2s edges Knowing a set of provider-free ASes, it is possible to draw economic inferences on AS paths Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
Transient AS paths 1 Inter-T1 routes All routes 0.1 CCDF 0.01 0.001 1 10 100 1000 10000 100000 1e+06 1e+07 Lifespan [s] Transient AS paths may be non compliant with the valley-free rule due to BGP misconfigurations that appear only during the BGP convergence process after a network failure Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
A time-aware tagging algorithm Inferences are merged together only if their lifespan ratio is larger than N MAG [1] E.Gregori, A.Improta, L.Lenzini, L.Sani, L.Rossi, “BGP and Inter-AS Economic Relationships” , in Proceedings of the 11th International IFIP TC-6 Conference on Networking (NETWORKING ’11), vol.2, pp. 54-67, 2011 Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
A filtering pre-phase to improve the algorithm Not every transient AS path is affected by BGP misconfigurations though... 0.3 All routes Routes lasted less than 60s Routes lasted less than 1h 0.25 Routes with inter-PF valleys 0.2 p(x) 0.15 0.1 0.05 0 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 AS path length (x) [2] E.Gregori, A.Improta, L.Lenzini, L.Sani, L.Rossi, “On Improving the Reliability of Inter-AS Economic Inferences Through an Hygiene Phase on BGP Data” , submitted to Computer Networks, 2013 Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
PhD Contribution 1 Inter-AS economic relationship inference from BGP data 2 Geographic AS path inference from BGP data 3 Quantification of BGP data completeness 4 Identification of ideal BGP feeders to improve current situation Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
BGP data and geography The global Internet is the result of the interconnection of multiple geographic networks ... ... but the analysis of the Internet from a global perspective may hide several regional characteristics e.g. Telstra (AS4637) is fundamental for the Australian connectivity, but represent just a non-stub in the global view Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
How to infer geographic information from BGP data “ An AS is a connected group of one or more IP prefixes run by one or more IP network operators which has a single and clearly defined routing policy ” (RFC 1930) For each AS: 1 We collect its IP prefixes from BGP data 2 We geolocate the AS by geolocating its prefixes (via Maxmind GeoLite Database) 3 We apply an heuristic to infer geographic AS paths and geographic topologies 41,796 out of 44,389 ASes result located only in one continent 40,347 out of 44,389 ASes result located only in one country [3] E.Gregori, A.Improta, L.Lenzini, L.Sani, L.Rossi, “Inferring Geography from BGP Raw Data” , in Proceedings of the 4th IEEE International Workshop on Network Science for Communication Networks (NETSCICOM ’12), pp. 208-213, 2012 [4] E.Gregori, A.Improta, L.Lenzini, L.Sani, L.Rossi, “Discovering the geographic properties of the Internet AS-level topology” , submitted to Networking Science, 2013 Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
PhD Contribution 1 Inter-AS economic relationship inference from BGP data 2 Geographic AS path inference from BGP data 3 Quantification of BGP data completeness 4 Identification of ideal BGP feeders to improve current situation Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
BGP Route Collector Status (Jan 2013) RouteViews RIS PCH BGPmon N. of RC 12 13 49 1 N. of feeders 116 309 936 33 Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
Feeder Contribution Only 152 feeders announce to the RCs their IPv4 full routing table Only 76 feeders announce to the RCs their IPv6 full routing table Alessandro Improta On the Analysis of the Internet from a Geographic and Economic Perspective via BGP Raw Data
Recommend
More recommend
