BGP and inter-AS economic relationships E. Gregori 1 , A. Improta 2 , - - PowerPoint PPT Presentation

BGP and inter-AS economic relationships

• E. Gregori1, A. Improta2,1, L. Lenzini2, L. Rossi1, L. Sani3

1 Institute of Informatics and Telematics, Italian National Research Council Pisa,

Italy

2 Information Engineering Department, University of Pisa, Italy 3 IMT Lucca, Institute for Advanced Studies, Lucca, Italy

Tuesday, May 10 - 2011

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 1 / 18

Paper goal

Development of an algorithm to infer economic relationships among ASes composing the Internet, that takes heavily into account the characteristics of exploited input data

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 2 / 18

Outline

1 Introduction and basilar concepts from previous works 2 Time-based tagging algorithm 3 Tagged topology analysis

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 3 / 18

Internet AS level topology

Internet: the largest collection of connected IP networks, belonging to various organizations (e.g. Universities, ISP, CDN) A group of network under the same organization running a the same routing protocol is an Autonomous System (AS), uniquely identified by a number (AS number - ASN) ASes can communicate to each other thanks to BGP, that allow them to build routes toward Internet networks A route is: “a unit of information that pairs a set of destinations with the attributes of a path to those destinations”

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 4 / 18

Data Gathering

Routes are stored in the RIB of each BGP router It is possible to study the Internet at the AS-level of abstraction exploting the AS path attribute AS paths are gathered by route-collectors deployed by RouteViews and RIPE-RIS projects 21,144,713 AS paths available (in October 2010)

36,437 ASes 116,671 connections

Usually the Internet is studied as an undirect graph, but this not take into account the different economic relationships existent among ASes

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 5 / 18

Economic Relationships

Typical economic relationships are: provider-customer: the customer pays the provider to reach all ASes that it cannot reach in other ways 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

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 6 / 18

Valley-free properties

Gao, in [1] pointed out that an AS should not transit traffic between two of its peers two of its providers a peer and a provider This leads to the valley-free property of AS paths:

A provider-customer edge can be followed by only provider-customer or sibling-to-sibling edges A peer-to-peer edge can be followed by only provider-customer

• r sibling-to-sibling edges

[1] Gao L., On inferring autonomous system relationships in the internet, IEEE/ACM Transactions on Networking, December 2001.

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 7 / 18

Tier-1 ASes

A Tier-1 AS is an AS that is able to reach all other Internet AS without recurring to any provider, i.e. a Tier-1 AS has no provider In [2], Oliveira et al. exploited valley-free property of AS paths and a list of Tier-1 ASes to build a new algorithm to infer AS relationships A non-Tier-1 AS should be able to reach all the Internet networks, thus there must exist at least one AS path including the considered AS and a Tier-1 AS The valley-free property and the presence of Tier-1 into an AS path, allow to infer the economic relationships between the ASes forming the path

[2] Oliveira et al., The (in)Completeness of the Observed Internet AS-level Structure, in IEEE/ACM Transactions on Networking, 2010.

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 8 / 18

Outline

1 Introduction and basilar concepts from previous works 2 Time-based tagging algorithm 3 Tagged topology analysis

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 9 / 18

Analysis of BGP data dynamics: methodology

Since ASes can change their decisions on how to reach ASes, AS paths can change during time The TIME information in update messages can be exploited to trace the dynamics of AS paths We consider as the lifespan of an AS path the max time span in which it was considered as active by the BGP

i.e. the longest time interval during which there is at least one active route that includes the considered AS path in its attributes.

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 10 / 18

CCDF Lifespan of gathered AS paths

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 100 101 102 103 104 105 106 107 P(X > x) Lifespan [s]

The CCDF highlights the presence of many short-lasting AS paths:

AS paths that appear after the activation of a backup connection Transient AS paths caused by BGP misconfigurations that appear only during the BGP convergence process after a network failure

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 11 / 18

Tagging algorithm

step A: Inference of all the possible economic relationships for each direct AS connection

direct means that (A,B) = (B,A)

It is based on the approach proposed by Oliveira et al. in [2]

The list of Tier-1 provided by Wikipedia has been exploited

For each tag is mantained the lifespan of the AS path used At the end of this step we have multiple (tag, lifespan) pairs for each connection

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 12 / 18

Tagging algorithm

step B: Inference of a single economic relationship for each direct AS connection All (tag, lifespan) pairs related to the same direct connection have to be merged

Find the max lifespan among each pair Merge only those pairs that have a comparable lifespan with the max, i.e. those do not differ more than N order of magnitude from the max Record the largest lifespan as the lifespan of the resulting tag

[A, B] [A, B] p2c p2p c2p s2s p2c p2c p2c s2s s2s p2p p2c p2p c2p s2s c2p s2s c2p c2p s2s s2s s2s s2s s2s s2s

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 13 / 18

Tagging algorithm

step C: Final tagging and two-way validation In order to have the economic relationship existing between AS A and AS B, the tags inferred for (A,B) and (B,A) connections have to be merged The approach used is the same as Step B, considering the different direction of connections, e.g. (A,B) = p2c and (B,A) = c2p have the same meaning

The merge is still based on lifespan, thus if the lifespans are not comparable, only the long-lasting tag affect the final tag

If there is a tag for both (A,B) and (B,A) and their lifespan is comparable, then the tag is said to be two-way validated

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 14 / 18

Outline

1 Introduction Basilar concepts from previous works 2 Time-based tagging algorithm 3 Tagged topology analysis

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 15 / 18

Results

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 100 101 102 103 104 105 106 107 P(X > x) Lifespan [s]

MIN lifespan of accepted (tag, lifespan) pairs in Step B

N=1 N=2 N=INF.

Results shows that if all the AS paths were used, the probability that transient AS paths affect the final results is larger The set of reliable tags is composed by the two-way validate tags, that represents

• nly the 4.5-6.4%

The precision of the algorithm suffers of the lack of information

NMAG Tag type Total One-way validated (%) 1 p2c 71841 69707 (97.0%) p2p 43398 41507 (95.6%) s2s 1378

• 2

p2c 72394 70034 (96.7%) p2p 42556 40716 (95.7%) s2s 1667

• ∞

p2c 74949 71642 (95.9%) p2p 39078 37235 (95.3%) s2s 2590

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 16 / 18

Stub ASes

NMAG Tag type Total Involving Stubs (%) 1 p2c 71841 53089 (73.8%) p2p 43398 12990 (29.9%) s2s 1378

• 2

p2c 72394 53331 (73.7%) p2p 42556 12748 (30.0%) s2s 1667

• ∞

p2c 74949 54180 (72.3%) p2p 39078 11898 (30.5%) s2s 2590

• The majority of the one-way validated connections are

involving to stub ASes A stub AS is an AS that does not transit traffic for any other AS Since their nature, stub ASes are likely to be customers of their neighbors

Considering this, the percentage of reliable tags rise to almost 56%

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 17 / 18

The End

Questions?

• E. Gregori, A. Improta, L. Lenzini, L. Rossi, L. Sani

BGP and inter-AS economic relationships 18 / 18