Computer Networks Zhenhai Duan Department of Computer Science 9/15/2011
Research Area Computer networks, in particular, Internet protocols, architectures, and systems Quality of Service (QoS) provisioning Internet inter-domain routing Internet systems security Overlay and peer-to-peer systems Network measurement Details and publications 2
A Few Projects that I will Discuss Packet scheduling algorithms Improving Internet inter-domain routing performance Controlling IP spoofing Detecting compromised machines (botnets) 3
QoS Provisioning on the Internet Current Internet provides a best-effort service No service guarantees in terms of bandwidth or end-to-end delay Many new applications require more stringent service guarantees VoIP and real-time video streaming Games Mission-critical applications Online financial transactions Power grid control system 4 Internet Can you hear me now?
5 Why current Internet cannot provide QoS guarantees? A number of factors (routing, architecture, etc) A key limitation is the First Come First Served (FCFS) packet scheduling algorithm used by routers
6 Two Fundamental Approaches to Designing New Packet Scheduling Round-robin packet scheduling algorithms Low complexity: O(1) Bad QoS performance: O(#flow)
7 Time stamp based fair queueing packet scheduling algorithms Emulating a single-flow system Time stamp based packet scheduling Compute and assign time stamps to each packet Scheduling based on time stamps Good performance: O(rate), largely independent of other flows High complexity: O(#flow) r C
More Scalable Packet Schedulers Hybrid round-robin and time-stamp based approach FRR IEEE INFOCOM 2005 IEEE ToC 2009 Core stateless packet schedulers VTRS, SETF, DETF ACM SIGCOMM 2000, IEEE ICNP 2001, IEEE ICCCN 2006 IEEE JSAC 2000, IEEE TPDS 2004, 2005 C 8
Internet Inter-Domain Routing Consists of large number of network domains (ASes) Each owns one or multiple network prefixes FSU campus network: /16 Intra-domain and inter-domain routing protocols Intra-domain: OSPF and IS-IS Inter-domain: BGP, a path-vector routing protocol BGP Used to exchange network prefix reachability information Network prefix, AS-level path to reach network prefix Path selection algorithm 9
10 BGP: an Example NLRI= /16 ASPATH=[0] /16 NLRI= /16 ASPATH=[10] NLRI= /16 ASPATH=[10] NLRI= /16 ASPATH=[210] NLRI= /16 ASPATH=[610] NLRI= /16 ASPATH=[610] NLRI= /16 ASPATH=[210] NLRI= /16 ASPATH=[7610] NLRI= /16 ASPATH=[4210] NLRI= /16 ASPATH=[3210] [3210]* [4210] [7610] NLRI= /16 ASPATH=[53210]
Performance Issues with BGP Instability At anytime, large number of BGP messages exchanged Slow convergence After a network failure event, it takes a long time for routing system to converge from one stable state to another stable state They are related, but not the same 11
Live BGP Updates Team Cymru BGPlay at RouteView 12
13 Network Dynamics Internet has about 38,600 ASes and 370,000 network prefixes (as of 09/03/2011) In a system this big, things happen all the time Fiber cuts, equipment outages, operator errors. Direct consequence on routing system Events may propagated through entire Internet Recomputing/propagating best routes Large number of BGP updates exchanged between ASes Effects on user-perceived network performance Long network delay Packet loss Even loss of network connectivity
Causes of BGP instability and long convergence Protocol artifacts of BGP Constraints of physical propagation Internet is a GLOBAL network [3210]* [4210] [7610] NLRI= /16 ASPATH=[57610] NLRI= /16 ASPATH=[54210] NLRI= /16 Withdrawal /16 14
Improving BGP stability and convergence BGP protocol artifacts EPIC: Carrying event origin in BGP updates Propagation delays on different paths Inter-domain failure vs. intra-domain failure Multi-connectivity between Ases Scalability and confidentiality IEEE INFOCOM 2005 Physical propagation constraints Transient failures TIDR: Localize failure events Build back-up paths IEEE GLOBECOM
Controlling IP Spoofing What is IP spoofing? Act to fake source IP address Used by many DDoS attacks Why it remains popular? Hard to isolate attack traffic from legitimate one Hard to pinpoint the true attacker Many attacks rely on IP spoofing cd ba s d c d s d s 16
Filtering based on Route A key observation Attackers can spoof source address, But they cannot control route packets take Requirement Filters need to compute best path from src to dst Filters need to know global topology info Not available in path-vector based Internet routing system cd ba s d s d s 17
Internet AS Relationship Consists of large number of network domains, Two common AS relationships Provider-customer Peering AS relationships determine routing policies A net effect of routing policies limit the number of routes between a pair of source and destination AS 2553 FSU AS FloridaNet AS 174 Cogent AS 3356 Level 3 AS2828 XO Comm AS Internet2 18
Topological Routes vs. Feasible Routes Topological routes Loop-free paths between a pair of nodes Feasible routes Loop-free paths between a pair of nodes that not violate routing policies cd ba s Topological routes s a d s b d s a b d s a c d s b a d s b c d s a b c d s a c b d s b a c d s b c a d Feasible routes s a d s b d cd ba s 19
Inter-Domain Packet Filter Identifying feasible upstream neighbors Instead of filtering based on best path, based on feasible routes Findings based on real AS graphs IDPFs can effectively limit the spoofing capability of attackers From 80% networks attackers cannot spoof source addresses IDPFs are effective in helping IP traceback All ASes can localize attackers to at most 28 Ases IEEE INFOCOM 2006, IEEE TDSC
Detecting Compromised Computers in Networks Botnet Network of compromised machines, with a bot program installed to execute cmds from controller, without owners knowledge. July 2009: Cyberattacks on government and commercial websites in US and South Korea About 50,000 compromised machines involved 21
Motivation and Problem Botnet becoming a major security issue Spamming, DDoS, identity theft sheer volume and wide spread 22
SPOT: Detecting Spam Zombies by Monitoring Outgoing Messages How to determine if a sending machine is compromised as s pass through SPOT sequentially Sequential probability ratio test (SPRT) IEEE INFOCOM 2009, IEEE TDSC (accepted) AB 23
Other Research Projects BGP Security ACM ASIACCS 2010 Spam filtering CEAS 2010, CEAS 2011 Detecting phishing s CEAS 2010 Security of anonymous networks Tor and Freenet 24
Thank you! Questions and comments? Details at my homepage 25
26 BGP Security Security relies on trust relationship among Ases Who owns which prefixes/how to reach Accidents (caused by human errors, not attacks) 24 Feb 2008, AS took Youtube’s /24 07 May 2005, AS 174 took Google’s /24 24 Dec 2004, Anatomy of a Leak: AS9121 ( 100K+ routes) 6 Apr 2001: C&W routing instability (f ull routing table announced) Check NANOG mailing list for more accidents Network prefix hijacking Origin spoofing, and path spoofing Existing solutions PKI-based secure BGP (S-BGP)
RBF: Region-Based BGP Update Filtering Two region granularities considered Country-level and RIR-level ACM ASIACCS