Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
An Integrated Approach to Improving Web Performance Lili Qiu Cornell University 1 Outline Motivation & Open Issues Solutions Study the workload of a busy Web server Optimize TCP performance for Web transfers Provision the content distribution networks Summary & Other Work 2 Motivation Web is the most dominant traffic in the Internet today Accounts for over 70% wide-area traffic Web performance is often unsatisfactory WWW – World Wide Wait Consequence: losing potential customers! Network congestion Overloaded Web server 3 Challenges in Providing Highly Efficient Web Services Protocol Inefficiency Workload characterization Workload Characterization Protocol inefficiency Infrastructure Provisioning The workload of busy Web sites is not well understood Mismatch between Web transfers and TCP protocol Infrastructure provisioning Current trend: Content Distribution Networks Problem: Where to place replicas? 4 Our Solutions Web Workload Characterization Improve protocol efficiency Study the workload of a busy Web server Optimize TCP startup performance for Web transfers Provision Web replication infrastructure Develop placement algorithms for content distribution networks (CDNs) 5 Part I Web Workload Characterization The Content and Access Dynamics of a Busy Web Site: Findings and Implications. Proceedings of ACM SIGCOMM 2000, Stockholm, Sweden, August 2000. (Joint work with V. N. Padmanabhan) 6 Motivation Solid understanding of Web workload is critical for designing robust and scalable systems Missing piece in previous work: workload of busy Web servers replica proxy Internet proxy Clients replica proxy Servers 7 Overview MSNBC server site Server logs a large news site consistently ranked among the busiest sites in the Web server cluster with 40 nodes 25 million accesses a day (HTML content alone) Period studied: Aug. – Oct. 99 & Dec. 17, 98 flash crowd HTTP access logs Content Replication System (CRS) logs HTML content logs Data analysis Content dynamics Access dynamics 8 Temporal Stability of File Popularity Methodology 17DEC98 - 18OCT99 01AUG99 - 18OCT99 17OCT99 - 18OCT99 Extent of overlap 1 0.8 0.6 0.4 Results 0.2 0 1 10 100 1000 # popular documents picked 10000 Consider the traces from a pair of days Pick the top n popular documents from each day Compute the overlap 100000 One day apart:significant overlap (80%) Two months apart: smaller overlap (20-80%) Ten months apart: very small overlap (mostly below 20%) The set of popular documents remains stable for days 9 Spatial Locality in Client Accesses Dec. 17, 1998 1.2 1 Fraction of requests shared Fraction of requests shared Normal Day 0.8 0.6 0.4 0.2 0 1 0.8 Trace 0.6 Random 0.4 0.2 0 0 10000 20000 30000 Domain ID 40000 50000 0 5000 10000 15000 20000 25000 30000 35000 Domain ID Domain membership is significant except when there is a “hot” event of global interest 10 Spatial Distribution of Client Accesses Cluster clients using network aware clustering [KW00] IP addresses with the same address prefix belongs to a cluster Top 10, 100, 1000, 3000 clusters account for about 24%, 45%, 78%, and 94% of the requests respectively A small number of client clusters contribute most of the requests. 11 The Applicability of Zipf-law to Web requests MSNBC Proxies Less popular servers 2 1.5 1 0.5 0 The Web requests follow Zipf-like distribution Request frequency 1/i, where i is a document’s ranking The value of is much larger in MSNBC traces 1.4 – 1.8 in MSNBC traces smaller or close to 1 in the proxy traces close to 1 in the small departmental server logs [ABC+96] Highest when there is a hot event 12 Impact of larger Percentage of Requests 1.2 1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 Percentage of Documents (sorted by popularity) 12/17/98 Server Traces 10/06/99 Proxy Traces 08/01/99 Server Traces Accesses in MSNBC traces are much more concentrated 90% of the accesses are accounted by Top 2-4% files in MSNBC traces Top 36% files in proxy traces (Microsoft proxies and the proxies studied in [BCF+99]) Top 10% files in small departmental server logs reported in [AW96] Popular news sites like MSNBC see much more concentrated accesses Reverse caching and replication can be very effective! 13 Part II Transport Layer Optimization for the Web Speeding Up Short Data Transfers: Theory, Architectural Support, and Simulation Results. Proceedings of NOSSDAV 2000 (Joint work with Yin Zhang and Srinivasan Keshav) 14 Motivation Characteristics of Web data transfers Short & bursty [Mah97] Use TCP Problem: Short data transfers interact poorly with TCP ! 15 TCP/Reno Basics Slow Start Congestion Avoidance Exponential growth in congestion window, Slow: log(n) round trips for n segments Linear probing of BW Fast Retransmission Triggered by 3 Duplicated ACK’s 16 Related Work P-HTTP [PM94] T/TCP [Bra94] Cache connection count, RTT TCP Control Block Interdependence [Tou97]: Reuses a single TCP connection for multiple Web transfers, but still pays slow start penalty Cache cwnd, but large bursts cause losses Rate Based Pacing [VH97] 4K Initial Window [AFP98] Fast Start [PK98, Pad98] Need router support to ensure TCP friendliness 17 Our Approach Directly enter Congestion Avoidance Choose optimal initial congestion window A Geometry Problem: Fitting a block to the service rate curve to minimize completion time 18 Optimal Initial cwnd Minimize completion time by having the transfer end at an epoch boundary. 19 Shift Optimization Minimize initial cwnd while keeping the same integer number of RTTs Before optimization: cwnd = 9 After optimization: cwnd = 5 20 Effect of Shift Optimization 21 TCP/SPAND Estimate network state by sharing performance information SPAND: Shared PAssive Network Discovery [SSK97] Internet Web Servers Performance gateway Directly enter Congestion Avoidance, starting with the optimal initial cwnd Avoid large bursts by pacing 22 Implementation Issues Scope for sharing and aggregation Collecting performance information Sliding window average Retrieving estimation of network state Performance reports, New TCP option, Windmill’s approach, … Information aggregation 24-bit heuristic network-aware clustering [KW00] Explicit query, active push, … Pacing Leaky-bucket based pacing 23 Opportunity for Sharing MSNBC: 90% requests arrive within 5 minutes since the most recent request from the same client network (using 24-bit heuristic) 24 Cost for Sharing MSNBC: 15,000-25,000 different client networks in a 5-minute interval during peak hours (using 24bit heuristic) 25 Simulation Results Methodology Performance Metric Download files in rounds Average completion time TCP flavors considered reno-ssr: Reno with slow start restart reno-nssr: Reno w/o slow start restart newreno-ssr: NewReno with slow start restart newreno-nssr: NewReno w/o slow start restart 26 Simulation Topologies 27 T1 Terrestrial WAN Link with Single Bottleneck 28 T1 Terrestrial WAN Link with Multiple Bottlenecks 29 TCP Friendliness 30 Summary TCP/SPAND significantly reduces latency for short data transfers 35-65% compared to reno-ssr / newreno-ssr 20-50% compared to reno-nssr / newreno-nssr Even higher for fatter pipes TCP/SPAND is TCP-friendly TCP/SPAND is incrementally deployable Server-side modification only No modification at client-side 31 Part III Provision Content Distribution Networks (CDNs) On the Placement of Web Server Replicas. To appear in INFOCOM'2001. (Joint work with V. N. Padmanabhan and G. M. Voelker) 32 Introduction to CDNs server server CDN server server server Content providers want to offer better service to their clients at lower cost Increasing deployment of content distribution networks (CDNs) Clients Content Providers Akamai, Digital Island, Exodus … Idea: a network of servers Features: Outsourcing infrastructure Improve performance by moving content closer to end users Flash crowd protection 33 Placement of CDN servers server Goal server CDN server server server minimize users’ latency or bandwidth usage Minimum K-median problem Select K centers to minimize the sum of assignment costs Clients Content Providers Cost can be latency or bandwidth or other metric we want to optimize NP-hard problem 34 Placement Algorithms Tree based algorithm [LGG+99] Random Assume the underlying topologies are trees, and model it as a dynamic programming problem O(N3M2) for choosing M replicas among N potential places Pick the best among several random assignments Hot spot Place replicas near the clients that generate the largest load 35 Placement Algorithms (Cont.) Greedy algorithm Greedy(N,M) { for I = 1 .. M { for each remaining replica R { cost[R] = cost after placing an additional replica at R } select the replica with the lowest cost } } Super Optimal algorithm Lagrangian relaxation + subgradient method 36 Simulation Methodology Network topology Randomly generated topologies Real Internet network topology AS level topology obtained using BGP routing data from a set of seven geographically dispersed BGP peers Web Workload Real server traces Using GT-ITM Internet topology generator MSNBC, ClarkNet, NASA Kennedy Space Center Performance Metric Relative performance: costpractical/costsuper-optimal 37 Simulation Results in Random Graph Topologies 38 Simulation Results in Real Internet Topologies 39 Effects of Imperfect Knowledge about Input Data Predict load using moving window average (a) Perfect knowledge about topology (b) Knowledge about Topology with a factor of 2 accurate 40 Summary First experimental study on placement of CDNs Knowledge about client workload and topology is crucial for provisioning CDNs The greedy algorithm performs the best The greedy algorithm is insensitive to noise Stay within a factor of 2 of the super-optimal when the salted error is a factor of 4 The hot spot algorithm performs nearly as well Within a factor of 1.1 – 1.5 of super-optimal Within a factor of 1.6 – 2 of super-optimal How to obtain inputs Moving window average for load prediction Using BGP router data to obtain topology information 41 Contributions Protocol Efficiency Workload characterization Protocol efficiency Workload Characterization Infrastructure Provisioning Study the workload of MSNBC web site Optimize TCP startup performance for Web transfers Infrastructure provisioning Develop placement algorithms for Content Distribution Networks 42 Other Work Available at http://www.cs.cornell.edu/lqiu/papers/papers.html Fast Firewall Implementations for Software and Hardware-based Routers. Submitted to ACM SIGMETRICS’2001. Integrating Packet FEC into Adaptive Voice Playout Buffer Algorithms on the Internet. Proceedings of IEEE INFOCOM'2000, Tel-Aviv, Israel, March 2000. On Individual and Aggregate TCP Performance. 7th International Conference on Network Protocols (ICNP'99), Toronto, Canada, October 1999. 43