wcw2004 overhaul

Upload: mahanbhavsar143

Post on 30-May-2018

231 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/9/2019 Wcw2004 Overhaul

    1/25

    Overhaul: Extending HTTP to

    Combat Flash CrowdsJay A. Patel & Indranil Gupta

    Distributed Protocols Research Group

    Department of Computer Science

    University of Illinois at Urbana-Champaign (UIUC)

    Urbana, Illinois, USA

  • 8/9/2019 Wcw2004 Overhaul

    2/25

  • 8/9/2019 Wcw2004 Overhaul

    3/25

    Distributed Protocols Research Group, Department of Computer Science, University of I ll inois at Urbana-Champaign 3

    Example: MSNBC

    MSNBC home pageDecember 14, 2003

  • 8/9/2019 Wcw2004 Overhaul

    4/25

    Distributed Protocols Research Group, Department of Computer Science, University of I ll inois at Urbana-Champaign 4

    Motivation Problem

    Unpredictable, yet frequent

    Brief period of time

    Thousand-fold increase in traffic

    Two nave solutions

    Overly insure on resources Shut down web site

  • 8/9/2019 Wcw2004 Overhaul

    5/25

    Distributed Protocols Research Group, Department of Computer Science, University of I ll inois at Urbana-Champaign 5

    Current Solutions Architectural Changes

    SEDA

    Capriccio ESI

    Protocol Modifications

    DHTTP

    Web Booster

    Cooperative Sharing

    Squirrel

    Kache Backslash

    BitTorrent

  • 8/9/2019 Wcw2004 Overhaul

    6/25

    Distributed Protocols Research Group, Department of Computer Science, University of I ll inois at Urbana-Champaign 6

    HTTP: RegularInteraction

    Client Server

    GET RequestResponse

    DocumentHeader

  • 8/9/2019 Wcw2004 Overhaul

    7/25

    Distributed Protocols Research Group, Department of Computer Science, University of I ll inois at Urbana-Champaign 7

    Overhaul: Overview Protocol change

    HTTP extension, no modification 5 new tags added, 1 slightly modified

    Backwards compatible Key concept: chunking

    Characteristic of the web applied to individual documents

    m chunks per document

    P2P distribution framework Voluntary

    Ad hoc, not DHT based

    Key benefit:parallel resource discovery

  • 8/9/2019 Wcw2004 Overhaul

    8/25

    Distributed Protocols Research Group, Department of Computer Science, University of I ll inois at Urbana-Champaign 8

    Overhaul: Design

    Client

    Server

    Client

    Client

    Client#1

    #2

    #4

    #3

    HTTP Request

    with Overhaul support tag

    Chunked Response

    with Overhaul headers

    Peers exchange chunks

    to fetch the complete document

    Ad hoc peer network

  • 8/9/2019 Wcw2004 Overhaul

    9/25

    Distributed Protocols Research Group, Department of Computer Science, University of I ll inois at Urbana-Champaign 9

    Details: Client/ServerInteraction Initial request by client

    Supports: Overhaul $port $speed

    Response by server in Overhaul mode

    ith chunk transmitted in sequential order

    Signatures of otherm-1 chunks for verification

    Initial Overhaul network membership list n most-recent Overhaul clients

    List maintained at server (updated with every request)

  • 8/9/2019 Wcw2004 Overhaul

    10/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 10

    Details: Peer Clients Interaction Clients contact other peer members

    To fetch remaining chunks

    To discover new peers

    Aggregate membership list by swapping information

    1-hop random walk discovery process

    Resource discovery

    Lookup documents on a busy Overhaul server

    Contact peers randomly on membership list

    INFO $host.tld

  • 8/9/2019 Wcw2004 Overhaul

    11/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 11

    Implementation Server

    Apache/2.0 HTTP server

    Module:mod_overhaul

    Client

    Java HTTP Proxy

    Cross platform Universal client support

  • 8/9/2019 Wcw2004 Overhaul

    12/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 12

    Testing Methodology: Server Server machine

    2.5 GHz AMD Athlon XP+

    1 GB RAM Client machine

    650 MHz Pentium III

    320 MB RAM

    Same network equipment

    25 concurrent fetches

    ApacheBench utility

  • 8/9/2019 Wcw2004 Overhaul

    13/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 13

    Results: Chunking (Fixed Size)Document: 10 KB

    Concurrency: 25

    Regular HTTP

    512-byte chunks

    2048-byte chunks

    Overhaul mode requires the server to send only a single chunk

  • 8/9/2019 Wcw2004 Overhaul

    14/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 14

    Results: Chunking (Maximum Count)

    Regular HTTP

    6 chunks

    12 chunks

    24 chunks

    Document: 50 KB

    Concurrency: 25

  • 8/9/2019 Wcw2004 Overhaul

    15/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 15

    Results: Overhaul vs. Regular

    Regular HTTP

    6 chunks

    12 chunks

    Concurrency: 25

    Minimum chunksize: 512-bytes

  • 8/9/2019 Wcw2004 Overhaul

    16/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 16

    Testing Methodology: Client Cluster of workstations

    25 homogenous PCs

    2.8 GHz Intel Pentium 4

    1 GB RAM

    Same network equipment

    Two experiments

    Concurrent: single document

    Staggered: multiple documents

  • 8/9/2019 Wcw2004 Overhaul

    17/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 17

    Results: Single Document Large document: 50

    KB (12 chunks)

    Server condition: 150-250 concurrent fetches

    + competition

    Overhaul requests:

    concurrently only using 24

    Overhaul-aware clients

    Regular

    requests

    Overhaul

    mode

    Fastest 1 sec 6 secs

    Slowest 32 secs 9 secs

    Average 9 secs 7 secs

    Server bandwidth usage in Overhaul mode: 1/12th of regular requests

  • 8/9/2019 Wcw2004 Overhaul

    18/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 18

    Results: Multiple Documents 8 documents: 110 KB total

    (12 chunks)

    Server condition: 150-250

    concurrent fetches +

    competition

    Overhaul requests staggered

    1st stage: 12 concurrent

    fetches, fetch all documents 2nd stage: 12 concurrent

    fetches, fetch index document

    only

    Regular

    requests

    Overhaul

    mode

    Fastest 1 sec 14 secs

    Slowest 28 secs

    Average 23 secs* 18 secs

    Server bandwidth usage in Overhaul mode : 1/18th of regular requests

    * indicates completed requests

  • 8/9/2019 Wcw2004 Overhaul

    19/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 19

    Limitations Both client and server must be Overhaul

    aware

    Requires critical mass to be maintained toremain effective

    n clients > m chunks

    More responsibilities for the client Possible security implications

  • 8/9/2019 Wcw2004 Overhaul

    20/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 20

    Conclusion Saves resources

    Bandwidth

    The bigger the crowd, the lower the per capita usage Response time

    Faster turnaround for both server and client

    Getting wide spread acceptance

    Marginal cost Protocol extension requires industry and

    standards push

  • 8/9/2019 Wcw2004 Overhaul

    21/25

  • 8/9/2019 Wcw2004 Overhaul

    22/25

  • 8/9/2019 Wcw2004 Overhaul

    23/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 23

    Heterogeneous Networks Problem

    Connections are

    heterogeneous

    Solutions

    Clustering of clients

    Super nodes

    Client

    Client

    Client

    Client

  • 8/9/2019 Wcw2004 Overhaul

    24/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 24

    Document Selection Only a partial set of

    documents are affected

    by a flash crowd Must implement

    selective Overhaul

    mode

    Automatic selection Active monitoring

    Server

    Large collection of documents

    reside on the server

    Documents fetched

    by a flash crowd

  • 8/9/2019 Wcw2004 Overhaul

    25/25

    Distributed Protocols Research Group, Department of Computer Science, University of Ill inois at Urbana-Champaign 25

    Dynamic Documents Flash crowds especially frequent during big

    events and news

    Characteristic: rapidly, changing data

    Solutions

    Time stamping

    Expiration of chunks Inter-network refresh from peers