Transcript darpa-agn
Network Support for Accountability Nick Feamster Georgia Tech Collaborative Response with David Andersen (CMU), Hari Balakrishnan (MIT), Scott Shenker (UC Berkeley/ICSI) Georgia Tech Response with Wenke Lee and Mustaque Ahamad Two Responses (“Double Duty”) • “Towards an Accountable Internet Architecture”, Andersen, Balakrishnan, Feamster, Shenker • “In-Band, Bottom-Up Support for Availability and Accountability”, Feamster, Lee, Ahamad 2 Internet Accountability What is it? • Mechanisms to identify, isolate, punish “bad behavior” • Distinct from accounting (cf. original Clark design goals) Why might the network need to support it? • Attacks on the routing system • Control over traffic • Tracking and mitigating malice – Spam – Botnets – Phishing 3 Facets of Internet Accountability • Source: defense against address forgery • Data-plane: identify faulty network elements • Control-plane: identify forged routing messages • Recourse to avoid faulty or malicious elements – Scalable network support for path diversity – Better mechanisms to curtail unwanted traffic 4 AIP: Accountable IP Hash of autonomous domain’s public key AD Globally valid endpoint identifier (cf. IPv6 CGA, HIP, etc.) EID • Refactoring of Internet addresses: AD:EID – AD: The autonomous domain of the host – EID: A globally unique endpoint identifier • Addresses are self-certifying • Why change addressing? – Forms the cornerstone of routing, forwarding, identity – Current address structure makes existing mechanisms clumsy – New structure retains simplicity at the network layer and above 5 Source Accountability • Problem: Sources can forge IP addresses – Can complete three-way handshakes (LAN spoofing) • Why it matters: Complicates filtering, blacklisting. – Spam campaigns regularly have 70% “fresh” IP addresses • Solution: Self-certification + Challenge/Response 6 Control-Plane Accountability • Problem: Routing messages can be forged • Why it matters: – Misconfiguration: AS 7007, ConEdison route leak – Malice: Spammers stealing address space • Solution: S-BGP-style attestations + self-cert – Interdomain routing and forwarding is on Ads – The AD is the public key and the address – Eliminates the need for a mapping between IP address space and organizations 7 Data-Plane Accountability • Problem: Network elements drop packets, fail, and otherwise give rise to poor performance • One Solution: In-Band Path Diagnosis New Shim Header IP Header Transport header Method Overview • Routers keep track of number of packets seen per flow • Each router stamps each packet with current flow counter value • If current counter value does not equal router’s expected packet count for that flow, router marks packet 8 “Cornerstone” Projects • In-Progress – Control-Plane: Implementation of control plane accountability (XORP) – Data-Plane: Implementation of in-band diagnosis (XORP/Quagga on VINI) • Imminent – Coping: Multiple ongoing efforts • Architectural mechanisms for supporting path choice • Scalable route diversity protocol that uses network virtualization (analysis, implementation on VINI) 9