iNET Seminar A Framework for Nation-Centric Classification and - - PowerPoint PPT Presentation

inet seminar
SMART_READER_LITE
LIVE PREVIEW

iNET Seminar A Framework for Nation-Centric Classification and - - PowerPoint PPT Presentation

iNET Seminar A Framework for Nation-Centric Classification and Observation of the Internet Sebastian Meiling Outline Introduction Methods Results Summary & Outlook 02/11/10 HAW Hamburg - iNET - Sebastian Meiling 2


slide-1
SLIDE 1

A Framework for Nation-Centric Classification and Observation of the Internet

Sebastian Meiling

iNET Seminar

slide-2
SLIDE 2

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 2

Outline

 Introduction  Methods  Results  Summary & Outlook

slide-3
SLIDE 3

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 3

Introduction

 Internet is a critical infrastructure

 Countries & governments  National & global economy

 Internet usage

 Communication  Information source  Market place  Business model  ...

slide-4
SLIDE 4

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 4

But ...

 The Internet is also subject to

 Censorship & surveillance  National laws & policies  Attacks & crime

 Upcoming questions

 How (well) is the connectivity of a country?  Who are important players in the Net?  What is the visualization and structure of a national

Internet topology graph

slide-5
SLIDE 5

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 5

Internet Building Blocks

 BGP inter-connects Autonomous Systems (AS)

 Each AS maintains one or more IP prefixes  An IP prefix consists of one or more IP blocks  An IP block is a range of IP addresses allocated by one

  • rganization (situated in a certain country)

 IP blocks are governed by regional registrars (RIRs)

 RIPE NCC, ARIN, APNIC, AFRINIC, and LACNIC

 Relationship between Internet topology and geography  IP blocks/addresses offer appropriate granularity to

discover country-wise (all) Internet players

slide-6
SLIDE 6

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 6

Outline

 Introduction  Methods  Results  Summary & Outlook

slide-7
SLIDE 7

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 7

From IP address to AS number

 Identify all IP blocks of a specific country

 query regional databases and BGP route monitors  for Germany: RIPE DB and RIPE RIS

 Map IP blocks to smallest enclosing prefix  Resolve IP prefixes to AS numbers  Retrieve additional AS information

 AS name, owning company, address data

slide-8
SLIDE 8

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 8

Toolchain

slide-9
SLIDE 9

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 9

Classification

 Topological category

 Based on classification by UCLA  Tier1, large and small ISP, stub

 Business branches/sectors

 keyword filter, manually refined  ISP, peering points, Traders, industry, research and

education, government, and others

slide-10
SLIDE 10

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 10

Problems

 Inconsistent or incomplete DB entries  Restricted access to certain information  Unresolvable IP blocks  Conflicting mappings between different DB  Special case: EU countries need additional

filtering on address data keywords

 Manual refinement of keywords

slide-11
SLIDE 11

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 11

Outline

 Introduction  Methods  Results  Summary & Outlook

slide-12
SLIDE 12

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 12

In numbers

 Identified 245524 German IP-blocks

 240237 embedded in 6278 IP-prefixes  Prefixes mapped to 1472 ASes  5286 IP-blocks remained unresolved (≈ 2%)

 IP-block vs. IP-prefix:

DE EU 6278 5243 1035 4395 947 936

  • ther

(our) IP-block approach IP-prefix approach (RIPE DB) IP-prefix approach (Team Cymru)

slide-13
SLIDE 13

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 13

Internet Exchange Points

slide-14
SLIDE 14

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 14

Betweenness

slide-15
SLIDE 15

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 15

Traders and financial services

slide-16
SLIDE 16

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 16

Outline

 Introduction  Methods  Results  Summary & Outlook

slide-17
SLIDE 17

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 17

Summary

 Tool chain to identify a nation-centric Internet

 demonstrated for Germany  special case: EU country

 Method starts from IP-blocks

 not from prefixes, like others  identifies more German ASes

 Sector filter

 sort ASes into business branches  analysis of AS relations

slide-18
SLIDE 18

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 18

Outlook

 Apply regular updates  Test and verify the methods and toolchain

against other countries

 (How) do the results differ?

slide-19
SLIDE 19

02/11/10 HAW Hamburg - iNET - Sebastian Meiling 19

References

[1] J.Karlin, S.Forrest and J.Rexford, “Nation-State Routing:

Censorship, Wiretapping, and BGP”, Tech.Rep. 2009

[2] M.Wählisch, S.Meiling, and T.C.Schmidt, “A Framework for

Nation-Centric Routing and Observation of the Internet”, CoNEXT'10 Student Workshop

[3] M.Wählisch, T.C.Schmidt, S.Meiling, Markus de Brün, and Thomas Häberlen, “Towards a Nation-Centric Understanding of the

Internet”, CoNEXT'10 Student Workshop

[4] B.Zhang, R.Liu, D.Massey, and L.Zhang, “Collecting the

Internet AS-Level Topology”, Comp.Comm.Rev. 2005