2. 2
Token Rings
• IBM Token Ring & IEEE 802.5
• FDDI
• Resilient Packet Ring IEEE 802.17…
•
IBM Token Rings are the most famous of this series
3. 3
About IBM Token Rings
• Network consists of a set of nodes connected in a
ring & data always flows in particular direction
• Each node receives frames from its upstream node
and transmits to its downstream node
• Token rings share two features with Ethernets
– Requires an algorithm that controls when each node is allowed
to transmit
– All nodes see all frames with destination saving a copy of it
4. 4
Tokens
• Special sequence of bits that circulates around
the ring
– 24 bits in case of 802.5
• When node with a frame to transmit sees the
token, it takes the token off the ring and instead
inserts its own frame on the network
– Frame is forwarded by all nodes while destination too makes a
copy of it and forwards it
– Packet is received again by sender which strips its frame off
the ring and reinserts the token
5. 5
Physical Properties
• Link failure in ring may disrupt transmission of the
whole network
• Electromechanical relay is used to overcome this
problem
• As long as the station is healthy, the relay is open
• When the stations stops powering the relay, it
gets closed and the station is removed from the
network
8. 8
Multi Station Access Unit… MSAU
• Several Electromechanical relays packed into a
single box
• It makes token rings look like a star topology
• Systems are plugged in and out of MSAU
• One of the main differences between 802.5 and
IBM Token Rings is that the latter requires MSAU
– The other difference however can be that that IBM Token
Rings can have up to 260 stations per ring while 802.5 supports
250
– Physical media for IBM is twisted pair but not defined for
802.5
10. 10
Token Ring Media Access Control
• Any station that wants to transmit data, has to
seize the token from the network
• Once the token is seized, transmission can be done
• In 802.5, seizing process involves simply modifying
1 bit in the token, and it acts as a preamble for
the subsequent data packet
11. 11
Data Transmission
• Each transmitted packet contains a destination
address
• As the packet flows past each node on the ring,
each node looks inside the packet to see if it is
the intended recipient
• If so, it copies the packet into its buffer as it
flows through the network adapter
• But it does not remove the packet form the
network
• The sending station has the responsibility of
removing the packet from the ring
• Other stations don’t harm the packet
12. 12
Token Holding Time
• Setting THT to infinity will be dangerous because
other devices wont have fair access to the ring
• Setting THT to one message will be silly
– If node has more packets to send, after every THT it would let
the token circulate around the whole ring before seizing it
again
• Hence it is set to 10ms
– A device should take care of the fact that before transmission
of a packet, whether it got sufficient THT left to transmit this
packet
13. 13
Token Rotation Time
• Time required by token to traverse the ring
• TRT < ActiveNodes * THT + RingLatency
• Active Nodes are the number of nodes that have
data to transmit and Ring Latency is the time a
packet takes to traverse the ring when no node
has data to send… Ideal
14. 14
Reliable Delivery
• A & C bits contained in frame status byte
• Both set to 0 by transmitter
• As the destination station sees the packet, it set
the A-bit to 1
• When the destination station copies the packet
into its buffer successfully, the C bit is set to 1
• At Sender again
– A=1 & C=1 means Successful Delivery
– A=1 & C=0 means Host is alive but frame could not be copied
15. 15
Priority
• Priority allows a particular host to access the
token
• Token contains a 3-bit priority field
• Token has certain priority n at any time
• Each host that wants to transmit a packet, has
to assign a priority to the data packet
• The host that holds the token, before releasing
the token sets its priority to the priority of
the data packet it received
16. 16
Token Ring Maintenance
• Monitor Station
• A monitor periodically announces its presence by
special control messages
• If these periodic messages stop arriving, the
monitor will be assumed to have failed
• And a new monitor is elected
– Each station that wants to be monitor sends a claim frame
– If the token circulates back to the sender, it will be the
monitor
– If more than one claim frames, break the tie by some algorithm
(say the highest address wins)
17. 17
Monitor
• Monitor is to take care of the health of the token
ring network
– Clocking… In case of Token
– Setting bits… In Case of Data Frame
• Reinserts a new token if token is not seen by
monitor after the interval
– (NumStations * THT) + RingLatency
• Checks for orphaned & Corrupted frames
o M-bit is set to 0 by transmitter and set to 1 as it passes the
monitor
o If a token passes the monitor with the m-bit already set to 1, it is
removed from the network