Application Architectures

Question 1

Application architecture

Answer 1

• Main categories of application architecture
  – Client-server
  – Peer-to-Peer (P2P)

Question 2

Client-server architecture

Answer 2

• Server:
  – An always-on host
  – Permanent IP address
  – Can use data centres for scaling
• Clients:
  – May be intermittently connected
  – May have dynamic IP addresses
  – Do not communicate directly with each other

Question 3

P2P architecture(s)

Answer 3

• No always-on server
• Arbitrary end systems communicate directly
• Peers request service from other peers, provide service in return to other peers
• Self-scalability – as well as making new service demands, new peers typically also bring new service capacity
• Peers are intermittently connected and IP addresses change routinely
  – Complex management

Question 4

P2P Networks: Napster

Answer 4

• Napster history: the rise
  – 1/99: Napster version 1
  – 5/99: company founded
  – 12/99: first lawsuits
  – 2000: 80 million users
• Napster history: the fall
  – Mid 2001: out of
  business due to lawsuits
  – Mid 2001: dozens of
  decentralized P2P
  alternatives
  – 2003: growth of pay
  services like iTunes

Question 5

How Napster worked

Answer 5

• Content location is centralised; but file transfer is decentralised
• Advantages:
  – Fast lookup time
• Disadvantages
  – Single point of failure for lookup
  – also: performance bottleneck
  – Copyright infringement

Question 6

P2P Networks: Gnutella

Answer 6

• Gnutella history
  – 2000: J. Frankel & T. Pepper released Gnutella
  – Soon after: many other clients (e.g., Morpheus, Limewire, Bearshare)
  – 2001: protocol enhancements, e.g., “supernodes/ultrapeers”
• Query flooding
  – Join: contact a few random(ish) nodes to become neighbours of
  – Search: I ask my neighbours, who ask their neighbours - if a node has it, it replies to me
  – Fetch: I get file directly from a node that claims to have it

Question 7

How Gnutella worked: Query flooding

Answer 7

• Advantages
  – Fully decentralised - no bottleneck
  – Distributed cost of searching
• Disadvantages
  – Potentially long search time
  – Prone to “free-loading”

Question 8

P2P Networks: Kazaa

Answer 8

• KaZaA history
  – 2001: created by Dutch company (Kazaa BV)
  – A single network called FastTrack was used by other clients as well
  – Eventually protocol changed so others could no longer use it
• Super-node hierarchy
  – Join: on startup, the client contacts a “supernode”
  – Publish: client sends a list of its available files to its super-node
  – Search: queries flooded among super-nodes
  – Fetch: get file directly from one or more peers

Question 9

How KaZaa worked

Answer 9

Question 10

P2P Networks: BitTorrent

Answer 10

• BitTorrent history
  – 2002: debuted by B. Cohen
• Emphasis on efficient fetching, not searching
  – Distribute same file to many peers
  – Single publisher, many downloaders
• Preventing “free-loading”
  – Incentives for peers to contribute

Question 11

BitTorrent Components

Answer 11

Question 12

How BitTorrent works

Answer 12

• A new peer joining a torrent:
  – registers with tracker to get list of peers, connects to subset of peers (“neighbours”)
  – has no chunks initially, but will accumulate them over time from other peers
• While downloading, peer uploads chunks to other peers
• Peer may change peers with whom it exchanges chunks
• Churn: peers may come and go
• Once peer has entire file, it may (selfishly) leave or (altruistically) remain in torrent

Question 13

P2P Sytems - Unstructured vs Structured

Answer 13

Question 14

Content Distribution (streaming)

Answer 14

• Challenge: how to stream content to hundreds of thousands of simultaneous users?
• Option 1: single, large “mega-server”
  – Single point of failure
  – Point of network congestion
  – Long path to distant clients
  – Multiple copies of video sent over outgoing link
• Option 2: store/serve multiple copies of videos at multiple geographically distributed sites (CDN)
  – Close to users
  – Resilient

Question 15

Content Distribution Networks

Answer 15

• CDN: stores copies of given content at CDN nodes
• Subscriber requests content from CDN
  
  • Is directed to nearby copy, retrieves content
  • may choose different copy if network is congested