Computer Science 328 Distributed Systems

1
Computer Science 328Distributed Systems

• Lecture 3
• NETWORKING

2
Networking Issues for DS

• ? Performance
• Latency (Lt)
• Propagation delay the time it takes for the
  first bit of a message to reach the destination.
• Processing delay the time it takes for the OS to
  process/send/ receive the message.
• Queueing delay the time it takes a message to be
  queued either at end hosts or intermediate nodes
  waiting for transmission.
• Data transfer rate (Tr)
• Bandwidth the total amount of information that
  can be transmitted over a given time.
• ?

packet delay Lt packet size / Tr,
3
Other Networking Issues for DS

• Scalability
• Reliability
• Security
• Mobility
• Quality of service
• Multicasting

4
Types of Networks

• Range Bandwidth Latency
• Local Area 1 - 2 km 10-1000Mbps 1-10
  ms
• Network (LAN)
• Wide Area worldwide 0.01- 600 100 - 500
  
• Network (WAN)
• Metro Area 2 - 50 1 - 150 10
• Network (MAN)
• Wireless LAN 0.15 - 1.5 2 11 5 - 20
• Wireless WAN worldwide 0.01 - 2
  100 - 500
• Internetwork worldwide 0.01 - 2
  100 - 500

5
Networking Principles

• Switching Schemes
• Broadcast
• Circuit Switching a physical circuit is
  established from the caller to the callee before
  communication takes place.
• Packet Switching (a.k.a. store-and-forward)
  packets that arrive at a node are first stored at
  the node and then forwarded toward their
  destinations.
• Frame Relay Similar to store-and-forward, but
  intermediate nodes route frames based on their
  first few bits. Frames as a whole are not stored
  at nodes but pass through the nodes as short
  streams of bits.

6
OSI 7-Layer Protocol Model

• Protocol
• Specifies the sequence of messages to be
  exchanged
• Specifies the format of data in messages

Message Sent
Message Received
Application
Presentation
Session
OSI Layered Protocol Model
Transport
Network
Data link
Physical
Communication Channel
7
Encapsulation of Upper-layer Data
8
Internet 5-Layer Model
9
Message Encapsulation -- TCP over an Ethernet
10
The Programmer's Conceptual View
TCP Transmission control protocol UDP User
datagram protocol
11
OSI 7-Layer Protocol Summary
12
OSI Reference Model

• Each layer performs a well defined function, and
  provides a well-defined service to the next
  higher layer. The protocol defines the operation
  that the peer modules use in exchanging messages
  so as to provide the required functions.
• The interface between layer n module and layer
  n-1 module at a node is precisely defined.
• A layer n module at one end communicates with its
  peer layer n module at the other end by passing a
  message into the layer n-1 module (along with
  necessary information for the lower layer module
  to generate headers).

13
Physical Layer

• Function Provides a virtual link, called virtual
  bit pipe, for transmitting a sequence of bits
  between any pair of nodes.
• Physical interface module (called modem) maps the
  incoming bits from the data link layer into
  signals appropriate for the channel, and at the
  receiving end, maps the signals back into bits.
• Interface module between the data link module and
  the modem (e.g., RS-232-C interface or X.21
  physical layer) is responsible for initiating
  data sending/receipt by providing separate wires
  between the two modules for control signals.

14
Data Link Layer

• Function Converts the unreliable bit pipe
  provided by the physical layer into a virtual
  communication link for sending packets error
  free.
• The sending DL module places some control bits
  called header at the beginning of each packet and
  some more overhead bits called trailer at the end
  of each packet, resulting in a longer string of
  bits called a frame.
• Some of the overhead bits perform error
  detection/correction, and some request
  retransmissions when error occurs.
• For multiaccess links (e.g., Ethernets), there is
  a need for a new sublayer, called the MAC
  sublayer, to arbitrate the access to the link so
  that each node can successfully transmit its
  frames without interference from the other nodes.

15
Network Layer

• The transport layer provides packets and control
  information (on how to handle the packets, e.g.,
  where the packets should be delivered) to the
  network layer. The network layer then uses the
  control information to generate the packet
  header.
• Function 1 Addressing
• Function 2 Routes packets from their sources
  through the network to their destinations
• Function 3 Congestion control

16
IP Addressing
7
24
Class A
0
Network ID
Host ID
14
16
Class B
1
0
Network ID
Host ID
21
8
28
Class D (multicast)
1
1
1
0
Multicast address
27
Class E (reserved)
1
1
1
1
unused
0
32-bit numeric identifier consisting of a network
identifier and a host identifier
17
Decimal Representation of Internet Addresses
octet 1
octet 2
octet 3
Range of addresses
Network ID
Host ID
1.0.0.0 to
Class A
1 to 127
0 to 255
0 to 255
0 to 255
127.255.255.255
Network ID
Host ID
128.0.0.0 to
Class B
128 to 191
0 to 255
0 to 255
0 to 255
191.255.255.255
Network ID
Host ID
192.0.0.0 to
Class C
0 to 255
0 to 255
1 to 254
192 to 223
223.255.255.255
Multicast address
Multicast address
224.0.0.0 to
Class D (multicast)
0 to 255
0 to 255
1 to 254
224 to 239
239.255.255.255
240.0.0.0 to
Class E (reserved)
0 to 255
0 to 255
1 to 254
240 to 255
255.255.255.255
18
IP Header
19
IP Packet Layout

• IP address allocation problem
• Network administrators usually overestimate
  future growth of their networks
• and request Class B address.
• Two existing solutions
• IP v6
• Classless interdomain routing (CIDR)

20
IPv6 Header Layout
0-8 time-insensitive traffic 8-15
time-sensitive traffic
Establishment of virtual circults
IPv6 addresses are 128 bits (16 bytes) long
sufficient to provide 7 x 1023 IP addresses per
square metre across the entire surface of the
Earth.
21
CIDR

• Ideas
• Allocate a bunch of contiguous class C addresses
  to a subnet requiring more than 255 addresses.
• Subdivide a Class B address space for allocation
  to multiple Class-C subnets.
• Approaches
• Add a mask field a bit pattern to select the
  portion of an IP address for representing the
  subnet address and for routing table comparison.
• For example, a Class C address space can be
  subdivided into 32 groups of 8 hosts, e.g.,
  138.37.95.248/29

8
21
22
Address Resolution between IP and Underlying
Networks

• Most hosts are attached to a LAN by an interface
  board that only understands LAN addresses. For
  example, every Ethernet board is equipped with a
  48-bit Ethernet address.
• The boards send and receive frames based on
  48-bit Ethernet addresses. They know nothing
  about the 32-bit IP addresses.
• Address Resolution Protocol (ARP) maps the IP
  addresses onto data link layer addresses (e.g.,
  Ethernet).

23
ARP

• Suppose host 1 wants to send a packet to host 2.
• Host 1 finds the IP address, 192.31.65.5, for
  host 2 by the Domain Name Service (DNS).
• Host 1 builds a packet with 192.31.65.5 and
  gives it to the IP software.
• The IP software looks at the address and knows
  the destination is on its own subnet.
• Host 1 broadcasts a packet onto the Ethernet
  asking Who owns IP address 192.31.65.5?''
• Host 2 will respond with its Ethernet address
  (E2).
• The IP software on host 1 builds an Ethernet
  frame addressed to E2, puts the IP packet
• in the payload field, and dumps it onto the
  Ethernet.
• The Ethernet board of hosts 2 detects this frame
  and causes an interrupt.

24
ARP

• The performance of ARP can be improved by caching
  the broadcast results.
• Host 1 can include its IP to Ethernet mapping the
  ARP packet so that host 2 does not have to make
  the inquiry if needed.
• What if host 1 wants to send packets to host 4?
• proxy ARP the CS router could be configured to
  respond to ARP requests for network 192.31.63.0.
• default Ethernet address host 1 can be
  configured to see that the destination is on a
  remote network and sends all packets to a default
  Ethernet address (E3).

25
Routing Algorithms

• Programmed in the network layer
• determine the route for each packet as it
  travels through the net
• dynamically update routing tables to reflect
  congestion, changes and failures.
• Two approaches
• link-state (e.g., OSPF)
• distance-vector (e.g., RIP)
• Routers send a summary of their table to every
  neighboring node, periodically, or upon change.
• receiving node updates its table (link cost).

26
Distance Vector Routing

• Also termed as distributed Bellman-Ford algorithm
  or Ford-Fulkerson algorithm, included in RIP
  (routing information protocol), AppleTalk, and
  Cisco routers.
• Each router maintains a table indexed by each
  router in the subnet and giving the best known
  distance to each router and which link to use to
  get there.
• Once every T seconds each router sends to each
  neighbor the table.

27
Distance Vector Routing

28
Routing Tables
29
Distance Vector Routing
30
Pseudo-Code for RIP
Send Each t seconds or when Tl changes, send Tl
on each non-faulty outgoing link. Receive
Whenever a routing table Tr is received on link
n for all rows Rr in Tr if (Rr.link not equal
n) Rr.cost Rr.cost 1 Rr.link n if
(Rr.destination is not in Tl) add Rr to Tl
// add new destination to Tl else for all rows Rl
in Tl if (Rr.destination Rl.destination and
(Rr.cost lt Rl.cost or Rl.link n)) Rl Rr //
Rr.cost lt Rl.cost remote node has better
route // Rl.link n remote node is more
authoritative
31
Link State Routing

• Each router must
• Discover the neighbors and learn their network
  addresses
• When a router is booted, it learns who its
  neighbors are by sending a special Hello packet
  on each point-to-point link.
• The router on the other end sends back a reply.
• Measure the delay or cost to each of its
  neighbors
• A router sends a special Echo packet over the
  link that the other end sends back immediately.
  By measuring the round-trip time, the sending
  router gets a reasonable delay estimate.
• Construct a packet telling all it has just
  learned.

32
Link State Routing

• A router broadcasts a link-state-advertisement
  (LSA) packet periodically or upon topology change.

33
Link State Routing

• Broadcast the LSA packet to all other routers in
  the subnet.
• Each packet contains a sequence number that is
  incremented for each new LSA packet sent.
• Each router keeps track of all the (source
  router, sequence) pairs it sees. When a new LSA
  packet comes in, it is checked against the pairs.
  If the new packet is new, it is forwarded on all
  the links except the one it arrived on.
• The age of each packet is included and is
  decremented once per second (per hop in reality).
  When the age hits zero, the information is
  discarded.
• Compute the shortest path to every other router
  with Dijkstras algorithm.

34
Congestion Control

• Is achieved by informing nodes along a route that
  congestion has occurred and asking them to reduce
  their packet transmission rate.
• Congestion information can be supplied by
• explicit transmission of special messages (called
  choked packets), or
• implementation of a specific transmission control
  protocol (e.g., TCP takes packet loss as an
  indication of congestion and will hence reduce
  its congestion window the number of packets
  allowed to be transmitted before receipt of a
  positive acknowledgement).