Overwiew Peer-to-Peer Networks Server based Networks Broadcast Networks Point-to-Point Networks Circuit Switching. Packet Switching. Message Switching. Networking Devices
What is Peer-to-Peer? A model of communication where every node in the network acts alike. As opposed to the Client-Server model, where one node provides services and other nodes use the services.
Advantages of P2P Network No central point of failure –E.g., the Internet and the Web do not have a central point of failure. –Most internet and web services use the client-server model (e.g. HTTP), so a specific service does have a central point of failure. Scalability –Since every peer is alike, it is possible to add more peers to the system and scale to larger networks.
Disadvantages of P2P Computing Decentralized coordination –How to keep global state consistent? –Need for distributed coherency protocols. All nodes are not created equal. –Computing power, bandwidth have an impact on overall performance. Programmability –As a corollary of decentralized coordination.
P2P Computing Applications File sharing Process sharing Collaborative environments
P2P File Sharing Applications Improves data availability Replication to compensate for failures. E.g., Napster, Gnutella, Freenet, KaZaA (FastTrack).
P2P Process Sharing Applications For large-scale computations Data analysis, data mining, scientific computing E.g., distributed.net, World-Wide Computer
P2P Collaborative Applications For remote real-time human collaboration. Instant messaging, virtual meetings, shared whiteboards, teleconferencing, tele- presence. E.g., talk, AOL Messenger, Yahoo! Messenger, Jabber, MS Netmeeting, NCSA Habanero, Games
P2P Topologies Centralized Ring Hierarchical Decentralized Hybrid
Centralized Manageable Coherent Extensible Fault Tolerant Secure Lawsuit-proof Scalable System is all in one place All information is in one place XNo one can add on to system XSingle point of failure Simply secure one host XEasy to shut down ?One machine. But in practice?
Ring Manageable Coherent Extensible Fault Tolerant Secure Lawsuit-proof Scalable Simple rules for relationships Easy logic for state XOnly ring owner can add Fail-over to next host As long as ring has one owner XShut down owner Just add more hosts
Hierarchical Manageable Coherent Extensible Fault Tolerant Secure Lawsuit-proof Scalable ½Chain of authority ½Cache consistency ½Add more leaves, rebalance ½Root is vulnerable XToo easy to spoof links XJust shut down the root Hugely scalable – DNS
Decentralized Manageable Coherent Extensible Fault Tolerant Secure Lawsuit-proof Scalable XVery difficult, many owners XDifficult, unreliable peers Anyone can join in! Redundancy XDifficult, open research No one to sue ?Theory – yes : Practice – no
Centralized + Ring Manageable Coherent Extensible Fault Tolerant Secure Lawsuit-proof Scalable Just manage the ring As coherent as ring XNo more than ring Ring is a huge win As secure as ring XStill single place to shut down Ring is a huge win Common architecture for web applications
Centralized + Decentralized Manageable Coherent Extensible Fault Tolerant Secure Lawsuit-proof Scalable XSame as decentralized ½Better than decentralized Anyone can still join! Plenty of redundancy XSame as decentralized Still no one to sue ?Looking very hopeful Best architecture for P2P networks?
Napster The P2P revolution is started. Central indexing and searching service File downloading in a peer-to-peer point-to- point manner.
Gnutella Peer-to-peer indexing and searching service. Peer-to-peer point-to-point file downloading using HTTP. A gnutella node needs a server (or a set of servers) to “start-up”… gnutellahosts.com provides a service with reliable initial connection points But introduces a new single point of failure!
The Gnutella protocol (v0.4) PING – Notify a peer of your existence PONG – Reply to a PING request QUERY – Find a file in the network RESPONSE – Give the location of a file PUSHREQUEST – Request a server behind a firewall to push a file out to a client.
Freenet Peer-to-peer indexing and searching service. Peer-to-peer file downloading. Files served use the same route as searches (not point-to-point) –Provides for anonymity.
KaZaA/Morpheus Hybrid indexing/searching model –Not centralized like Napster, not decentralized like Gnutella. Peer-to-peer file downloading using HTTP. –“SmartStream” for incomplete file downloads. –“FastStream” for partial file downloads. “SuperNodes” elected dynamically if sufficient bandwidth and processing power – hybrid topology model. A central server keeps user registrations, logs usage, and helps bootstrapping peer discovery.
Client host Server host Client host Server Based Networks
Specialization: –Clients specialize in user interface –Servers specialize in managing data and application logic Sharing: –Many clients can be supported by few servers –Often data and logic are shared among applications and users
Client/server Peer-to-peer Server “I want to access some information” “I want to collaborate with my colleague” Client
Distinctions Client-server –Asymmetric relationship –Client predominately makes requests, server makes replies Peer-to-peer –Symmetric relationship
Client Server Client Email client sends message to server Message is stored on POP server Later, recipient’s email client retrieves message from server Email application
Client Server Client Chat clients send user’s typing to server Chat server aggregates typing from all users and sends to all clients Other user’s clients display aggregated typing from chat server Chat application
Presentation Application logic Shared data Local-area network Note: many clients per application server, several application servers per data server Three-tier client/server
Client Web browser Web server Application logic Databases and DBMS Common gateway interchange Host architecture Application partition
Broadcast Network: Shared Communication Medium –Shared Medium can be a set of wires Inside a computer, this is called a bus All devices simultaneously connected to devices –Originally, Ethernet was a broadcast network All computers on local subnet connected to one another –More examples (wireless: medium is air): cellular phones, GSM GPRS, EDGE, CDMA 1xRTT, and 1evDO Broadcast Networks Memory Processor I/O Device I/O Device I/O Device
Broadcast Networks Details Delivery: When you broadcast a packet, how does a receiver know who it is for? (packet goes to everyone!) –Put header on front of packet: [ Destination | Packet ] –Everyone gets packet, discards if not the target –In Ethernet, this check is done in hardware No OS interrupt if not for particular destination –This is layering: we’re going to build complex network protocols by layering on top of the packet Header (Dest:2) Body (Data) Message ID:1 (ignore) ID:2 (receive) ID:4 (ignore) ID:3 (sender)
Broadcast Network Arbitration Arbitration: Act of negotiating use of shared medium –What if two senders try to broadcast at same time? –Concurrent activity but can’t use shared memory to coordinate! Aloha network (70’s): Blind broadcast, with checksum at end of packet. If received correctly (not garbled), send back an acknowledgement. If not received correctly, discard. –Sender waits for a while, and if doesn’t get an acknowledgement, re-transmits. –If two senders try to send at same time, both get garbled, both simply re-send later. –Problem: Stability: what if load increases? More collisions less gets through more resent more load… More collisions… Unfortunately: some sender may have started in clear, get scrambled without finishing
Carrier Sense, Multiple Access/Collision Detection Ethernet (early 80’s): first practical local area network –It is the most common LAN for UNIX, PC, and Mac –Use wire instead of radio, but still broadcast medium Key advance was in arbitration called CSMA/CD: Carrier sense, multiple access/collision detection –Carrier Sense: don’t send unless idle Don’t mess up communications already in process –Collision Detect: sender checks if packet trampled. If so, abort, wait, and retry. –Backoff Scheme: Choose wait time before trying again How long to wait after trying to send and failing? –What if everyone waits the same length of time? Then, they all collide again at some time! –Must find way to break up shared behavior with nothing more than shared communication channel Adaptive randomized waiting strategy: –Adaptive and Random: First time, pick random wait time with some initial mean. If collide again, pick random value from bigger mean wait time. Etc. –Randomness is important to decouple colliding senders
Point-to-point networks Why have a shared bus at all? Why not simplify and only have point-to-point links + routers/switches? –Didn’t used to be cost-effective –Now, easy to make high-speed switches and routers that can forward packets from a sender to a receiver. Point-to-point network: a network in which every physical wire is connected to only two computers Switch: a bridge that transforms a shared-bus (broadcast) configuration into a point-to-point network. Router: a device that acts as a junction between two networks to transfer data packets among them. Router Internet Switch
Point-to-Point Networks Advantages: –Higher link performance Can drive point-to-point link faster than broadcast link since less capacitance/less echoes (from impedance mismatches) –Greater aggregate bandwidth than broadcast link Can have multiple senders at once –Can add capacity incrementally Add more links/switches to get more capacity –Better fault tolerance Disadvantages: –More expensive than having everyone share broadcast link Examples –ATM (asynchronous transfer mode) The first commercial point-to-point LAN Inspiration taken from telephone network –Switched Ethernet Same packet format and signaling as broadcast Ethernet, but only two machines on each ethernet.
CIRCUIT-SWITCHED NETWORKS CIRCUIT-SWITCHED NETWORKS A circuit-switched network consists of a set of switches connected by physical links. A connection between two stations is a dedicated path made of one or more links. However, each connection uses only one dedicated channel on each link. Each link is normally divided into n channels by using FDM or TDM.
In circuit switching, the resources need to be reserved during the setup phase; the resources remain dedicated for the entire duration of data transfer until the teardown phase. Circuit switching takes place in Physical layer. Data transfer between the two nodes is a continuous flow. No addressing is involved during data transfer.
As a trivial example, let us use a circuit-switched network to connect eight telephones in a small area. Communication is through 4-kHz voice channels. We assume that each link uses FDM to connect a maximum of two voice channels. The bandwidth of each link is then 8 kHz. Figure 8.4 shows the situation. Telephone 1 is connected to telephone 7; 2 to 5; 3 to 8; and 4 to 6. Of course the situation may change when new connections are made. The switch controls the connections. Example
As another example, consider a circuit-switched network that connects computers in two remote offices of a private company. The offices are connected using a T-1 line leased from a communication service provider. There are two 4 × 8 (4 inputs and 8 outputs) switches in this network. For each switch, four output ports are folded into the input ports to allow communication between computers in the same office. Four other output ports allow communication between the two offices. Example
DATAGRAM NETWORKS DATAGRAM NETWORKS In data communications, we need to send messages from one end system to another. If the message is going to pass through a packet-switched network, it needs to be divided into packets of fixed or variable size. The size of the packet is determined by the network and the governing protocol. In a packet-switched network, there is no resource reservation; resources are allocated on demand.
A datagram network with four switches (routers)
Routing table in a datagram network A switch in a datagram network uses a routing table that is based on the destination address. The destination address in the header of a packet in a datagram network remains the same during the entire journey of the packet. Switching in the Internet is done by using the datagram approach to packet switching at the network layer.
VIRTUAL-CIRCUIT NETWORKS VIRTUAL-CIRCUIT NETWORKS A virtual-circuit network is a cross between a circuit- switched network and a datagram network. It has some characteristics of both.
Setup acknowledgment in a virtual-circuit network
In virtual-circuit switching, all packets belonging to the same source and destination travel the same path; but the packets may arrive at the destination with different delays if resource allocation is on demand. Switching at the data link layer in a switched WAN is normally implemented by using virtual-circuit techniques.
Message Switching A store-and-forward network where the block of transfer is a complete message. Since messages can be quite large, this can cause: –buffering problems –high mean delay times