Brief Overview of Academic Research on P2P

Download Report

Transcript Brief Overview of Academic Research on P2P

Brief Overview of
Academic Research on
P2P
Pei Cao
Relevant Conferences





IPTPS (International Workshop on Peer-to-Peer
Systems)
ICDCS (IEEE Conference on Distributed
Computer Systems)
NSDI (USENIX Symposium on Network
System Design and Implementation)
PODC (ACM Symposium on Principles of
Distributed Computing)
SIGCOMM
Areas of Research Focus

Gnutella-Inspired


BitTorrent-Inspired



The “Directory Service” Problem
The “File Distribution” Problem
P2P Live Streaming
P2P and Net Neutrality
Gnutella-Inspired
Research Studies
The Applications and The Problems


Napster, Gnutella, KaZaa/FastTrak, Skype
Look for a particular content/object, and find
which peer has it  the “directory service”
problem


Challenge: how to offer a scalable directory
service in a fully decentralized fashion
Arrange direct transfer from the peer  the
“punch a hole in the firewall” problem
Decentralized Directory Services

Structured Networks

DHT (Distributed Hash Tables)
Very active research areas from 2001 to 2004
 Limitation: lookup by keys only


Multi-Attribute DHT


Limited support for query-based lookup
Unstructured Networks

Various improvements to basic flooding based
schemes
What Is a DHT?

Single-node hash table:
key = Hash(name)
put(key, value)
get(key) -> value

How do I do this across millions of hosts on the
Internet?

Distributed Hash Table
Distributed Hash Tables







Chord
CAN
Pastry
Tapastry
Symphony
Koodle
etc.
The Problem
N2
N1
Put (Key=“title”
Value=file data…)
Publisher
Internet
N4
• Key
Placement
• Routing to find key
N5
N3
?
N6
Client
Get(key=“title”)
Key Placement

Traditional hashing
Nodes numbered from 1 to N
 Key is placed at node (hash(key) % N)


Why Traditional Hashing have problems
Consistent Hashing: IDs

Key identifier = SHA-1(key)
Node identifier = SHA-1(IP address)
SHA-1 distributes both uniformly

How to map key IDs to node IDs?


Consistent Hashing: Placement
Key 5
Node 105
K5
N105
K20
Circular 7-bit
ID space
N32
N90
K80
A key is stored at its successor: node with next higher ID
Basic Lookup
N120
N10 “Where is key 80?”
N105
“N90 has K80”
K80 N90
N60
N32
“Finger Table” Allows log(N)-time
Lookups
¼
1/8
1/16
1/32
1/64
1/128
N80
½
Finger i Points to Successor of n+2i
N120
112
¼
1/8
1/16
1/32
1/64
1/128
N80
½
Lookups Take O(log(N)) Hops
N5
N10
K19
N20
N110
N99
N32 Lookup(K19)
N80
N60
Chord Lookup Algorithm
Properties


Interface: lookup(key)  IP address
Efficient: O(log N) messages per lookup




N is the total number of servers
Scalable: O(log N) state per node
Robust: survives massive failures
Simple to analyze
Related Studies on DHTs

Many variations of DHTs
Different ways to choose the fingers
 Ways to make it more robust
 Ways to make it more network efficient


Studies of different DHTs


What happens when peers leave aka churns?
Applications built using DHTs
Tracker-less BitTorrent
 Beehive --- a P2P based DNS system

Directory Lookups: Unstructured
Networks

Example: Gnutella

Support more flexible queries

Typically, precise “name” search is a small portion of all
queries
Simplicity
 High resilience against node failures


Problems: Scalability

Flooding  # of messages ~ O(N*E)
Flooding-Based Searches
1
3
2
5
6
4
7
8
. . . . . . . . . . . .


Duplication increases as TTL increases in
flooding
Worst case: a node A is interrupted by N * q *
degree(A) messages
Problems with Simple TTLBased Flooding

Hard to choose TTL:
For objects that are widely present in the network,
small TTLs suffice
 For objects that are rare in the network, large TTLs
are necessary


Number of query messages grow exponentially
as TTL grows
Idea #1: Adaptively Adjust TTL

“Expanding Ring”


Multiple floods: start with TTL=1; increment TTL
by 2 each time until search succeeds
Success varies by network topology
Idea #2: Random Walk

Simple random walk


takes too long to find anything!
Multiple-walker random walk
N agents after each walking T steps visits as many
nodes as 1 agent walking N*T steps
 When to terminate the search: check back with the
query originator once every C steps

Flexible Replication


In unstructured systems, search success is
essentially about coverage: visiting enough nodes
to probabilistically find the object =>
replication density matters
Limited node storage => what’s the optimal
replication density distribution?
In Gnutella, only nodes who query an object store it
=> ri  pi
 What if we have different replication strategies?

Optimal ri Distribution


Goal: minimize ( pi/ ri ), where  ri =R
Calculation:

introduce Lagrange multiplier , find ri and  that
minimize:
( pi/ ri ) +  * ( ri - R)
=>
 - pi/ ri2 = 0 for all i
=>
ri   pi
Square-Root Distribution


General principle: to minimize ( pi/ ri ) under
constraint  ri =R, make ri proportional to
square root of pi
Other application examples:
Bandwidth allocation to minimize expected
download times
 Server load balancing to minimize expected request
latency

Achieving Square-Root
Distribution

Suggestions from some heuristics



Store an object at a number of nodes that is proportional to
the number of node visited in order to find the object
Each node uses random replacement
Two implementations:


Path replication: store the object along the path of a
successful “walk”
Random replication: store the object randomly among nodes
visited by the agents
KaZaa



Use Supernodes
Regular Nodes : Supernodes = 100 : 1
Simple way to scale the system by a factor of
100
BitTorrent-Inspired
Research Studies
Modeling and Understanding
BitTorrent

Analysis based on modeling

View it as a type of Gossip Algorithm





Usually do not model the Tit-for-Tat aspects
Assume perfectly connected networks
Statistical modeling techniques
Mostly published in PODC or SIGMETRICS
Simulation Studies



Different assumption of bottlenecks
Varying details of the modeling of the data transfer
Published in ICDCS and SIGCOMM
Studies on Effect of BitTorrent on
ISPs

Observation: P2P contributes to cross-ISP
traffic


SIGCOMM 2006 publication on studies in Japan
backbone traffic
Attempts to improve network locality of
BitTorrent-like applications
ICDCS 2006 publicatoin
 Academic P2P file sharing systems


Bullet, Julia, etc.
Techniques to Alleviate the “Last
Missing Piece” Problem

Apply Network Coding to pieces exchanged between
peers


Pablo Rodriguez Rodriguez, Microsoft Research (recently
moved to Telefonica Research)
Use a different piece-replication strategy



Dahlia Makhi, Microsoft Research
“On Collaborative Content Distribution Using MultiMessage Gossip”
Associate “age” with file segments
Network Coding

Main Feature
Allowing intermediate nodes to encode packets
 Making optimal use of the available network
resources


Similar Technique: Erasure Codes

Reconstructing the original content of size n from
roughly a subset of any n symbols from a large
universe of encoded symbols
Network Coding in P2P: The Model

Server
Dividing the file into k blocks
 Uploading blocks at random to different clients


Clients (Users)
Collaborating with each other to assemble the blocks
and reconstruct the original file
 Exchanging information and data with only a small
subset of others (neighbors)
 Symmetric neighborhood and links

Network Coding in P2P
Assume a node with blocks B1, B2, …, Bk
 Pick random numbers C1, C2, …, Ck
 Construct new block
E = C1 * B1 + C2 * B2 + … + Ck * Bk
 Send E and (C1, C2, …, Ck) to neighbor
 Decoding: collect enough linearly independent E’s,
solve the linear system


If all nodes pick vector C randomly, chances are high that
after receiving ~K blocks, can recover B1 through Bk
P2P Live Streaming
Motivations

Internet Applications:


PPLive, PPStream, etc.
Challenge: QoS Issues

Raw bandwidth constraints


Example: PPLive utilizes the significant bandwidth
disparity between “Univeristy nodes” and “Residential
nodes”
Satisfying demand of content publishers
P2P Live Streaming Can’t Stand on
Its Own

P2P as a complement to IP-Multicast


P2P as a way to reduce server load


Used where IP-Multicast isn’t enabled
By sourcing parts of streams from peers, server load
might be reduced by 10%
P2P as a way to reduce backbone bandwidth
requirements

When core network bandwidth isn’t sufficient
P2P and NetNeutrality
It’s All TCP’s Fault


TCP: per-flow fairness
Browsers
2-4 TCP flows per web server
 Contact a few web servers at a time
 Short flows


P2P applications:
Much higher number of TCP connections
 Many more endpoints
 Long flows

When and How to Apply Traffic
Shaping


Current practice: application recognition
Needs:
An application ignostic way to trigger the traffic
shaping
 A clear statement to users on what happens
