Download presentation
Presentation is loading. Please wait.
1
ECE 544 Protocol Design Project 2016
Jamaal Alasadi Jose Wilfredo Aleman Espinoza Ioannis Paraskevakos
2
Network Architecture & Topology Assumptions
Unreliable network with packet loss prob. per link of p Each end node is attached to only one router All links have same characteristics (i.e. hop cost 1 and same MTU 1500 Bytes) Small maximum number of nodes ( ), hosts (256)
3
Network Architecture & Topology Assumptions
Mobility is supported using the same logic to deal with packet loss 4 links per router Fixed Static Address A packet, if received, is received correctly, The length of the packet is not fixed, it can send packets less than 1500 Protocol runs over the layer of Ethernet Summary of protocol design (packet structure, control plane for protocol, data plane for forwarding, key algorithms for mcast) Mobility means that an end node can move from one part of the network to another freely, but it has to be offline while moving. 3
4
Protocol Concept Overall concept of the proposed protocol
Provided service: File transfer The protocol provides a reliable delivery of the packets 16 bits for addresses (the less significant bit of the second Byte will represent the type of host/node, Router=1, Host=0) 1500 Bytes is max length of the packet 4
5
Syntax – Packet Formats (1/2)
Type #Dest Destination Address 1 Destination Address 2 Destination Address 3 Source Address Rand Gen Mess # (RGMN) F PL DATA
6
Syntax – Packet Formats (2/3)
Control Packets ARQ Used: Stop and wait for multicast and Go back N for unicast transmittion Control messages used: Acks: The header with the same message number and type equal to 2 Also used for discovering end-nodes (terminals) An Ack has to be sent even to a hello message from a router Time out after half a 0.5 mS We can use two methods stop and wait , and Go Back N , but we suggest Go Back N for unicast and stop and wait for multicast 100 routers in line 50μS for the message to reach, 120 μS for receiving all the packet and 50μS for the ack to return + plus additional time for processing.
7
Syntax – Packet Formats (3/3)
Control Packets Control messages used: Hello Messages: Used for discovery A Hello message can be send either by an end-node when it is connected to the network or by routers whenever there is a change in the routing table. The Hello Message from a End-Node is a packet header of type broadcast. The Receiving router is responsible to update its Routing table and inform its peers of the new Node The Hello message from a Router is a broadcast message the contains as data the Router's routing table. Explain the difference between the end node hello message and the router’s hello message. 7
8
Semantics Type: The type of the packet
0: Unicast 1: Multicast 2: Acknowledge 3: Broadcast #Dest: Number of receivers for a packet 0: Broadcast 1: One Receiver 2: Two Receivers 3: Three Receivers F(ragmentation): Shows if the packet is part of a fragmented message 0 – Single or Last Packet of a transmission 1 – First or Intermediate packet of a fragmented message RGMN: Random Number. It is used as a packet ID. Must be returned with Ack Packet length: 2 Bytes Addressing Scheme:MSB 0 End Node,MSB 1 router In the case multiple packet with a fragmented message the first received ID is the seed for the RNG.
9
Algorithms -Rooting: Distance Vector - Multicast Forwarding:
For all Destinations combinations: Find the port with the less combined cost. Otherwise copy accordingly. Combined Cost = maxport{dest_cost+1: every dest in port} *In case of a tie select port with the smallest value Do we need to know the paths from all ports?
10
Network 1: 1 - Identifying the next hops Minimum Cost Combined Path
11
Network 1: 2 - Identifying the Minimum Cost Combined Path
12
Network 2: 1 - Identifying the next hops
13
Network 2: 2 - Identifying the Minimum Cost Combined Path
14
Summary Key protocol design features (recap) Implementation complexity
Small Header overhead. 12 Bytes Reliability Single packet format for all types of messages. Reduces complexity in packet decoding. Implementation complexity The routing algorithm traverses through all end-nodes that are in the network. Making the overall computation complexity O(n), where n the number of end-nodes. For Storage we need an Array for storing all the end nodes and the port and path cost to each node and a second array to keep the table for the multicast messages. Finally, each end-node should have about 47KB for buffering packets Latency of 100 Mbps ethernet is 120us. So achieved throughput is Effective transfer size (DATA portion of the packet)/Transmit Time( Latency + 1-p in case we lose a packet and have to retransmit it and 99 Mbps, because we are using 100Mbps ethernet cable and 99% of the maximum packet is for data. Maximum Ethernet cable length 100m. Performance ….justify achieved throughput/overhead/delay based on network examples Achieved Maximum Throughput is (1-p)*99 Mbps Propagation delay between two routers/nodes 0.5μS.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.