Almost all collisions can be avoided in CSMA/CD but they can still occur during the contention period. The collision during the contention period adversely affects the system performance, this happens when the cable is long and length of packet are short. This problem becomes serious as fiber optics network came into use. Here we shall discuss some protocols that resolve the collision during the contention period.
- Bit-map Protocol
- Binary Countdown
- Limited Contention Protocols
- The Adaptive Tree Walk Protocol
Pure and slotted Aloha, CSMA and CSMA/CD are Contention based Protocols:
- Try-if collide-Retry
- No guarantee of performance
- What happen if the network load is high?
Collision Free Protocols:
- Pay constant overhead to achieve performance guarantee
- Good when network load is high
1. Bit-map Protocol:
Bit map protocol is collision free Protocol. In bitmap protocol method, each contention period consists of exactly N slots. If any station has to send frame, then it transmits a 1 bit in the corresponding slot. For example, if station 2 has a frame to send, it transmits a 1 bit to the 2nd slot.
In general, Station 1 Announce the fact that it has a frame questions by inserting a 1 bit into slot 1. In this way, each station has complete knowledge of which station wishes to transmit. There will never be any collisions because everyone agrees on who goes next. Protocols like this in which the desire to transmit is broadcasting for the actual transmission are called Reservation Protocols.
Bit Map Protocol fig (1.1)
For analyzing the performance of this protocol, We will measure time in units of the contention bits slot, with a data frame consisting of d time units. Under low load conditions, the bitmap will simply be repeated over and over, for lack of data frames. All the stations have something to send all the time at high load, the N bit contention period is prorated over N frames, yielding an overhead of only 1 bit per frame.
Generally, high numbered stations have to wait for half a scan before starting to transmit low numbered stations have to wait for half a scan(N/2 bit slots) before starting to transmit, low numbered stations have to wait on an average 1.5 N slots.
2. Binary Countdown:
Binary countdown protocol is used to overcome the overhead 1 bit per binary station. In binary countdown, binary station addresses are used. A station wanting to use the channel broadcast its address as binary bit string starting with the high order bit. All addresses are assumed of the same length. Here, we will see the example to illustrate the working of the binary countdown.
In this method, different station addresses are read together who decide the priority of transmitting. If these stations 0001, 1001, 1100, 1011 all are trying to seize the channel for transmission. All the station at first broadcast their most significant address bit that is 0, 1, 1, 1 respectively. The most significant bits are read together. Station 0001 see the 1 MSB in another station address and knows that a higher numbered station is competing for the channel, so it gives up for the current round.
Other three stations 1001, 1100, 1011 continue. The next station at which next bit is 1 is at station 1100, so station 1011 and 1001 give up because there 2nd bit is 0. Then station 1100 starts transmitting a frame, after which another bidding cycle starts.
Binary Countdown fig (1.2)
3. Limited Contention Protocols:
- Collision based protocols (pure and slotted ALOHA, CSMA/CD) are good when the network load is low.
- Collision free protocols (bitmap, binary Countdown) are good when load is high.
- How about combining their advantages :
- Behave like the ALOHA scheme under light load
- Behave like the bitmap scheme under heavy load.
4. Adaptive Tree Walk Protocol:
- partition the group of station and limit the contention for each slot.
- Under light load, everyone can try for each slot like aloha
- Under heavy load, only a group can try for each slot
- How do we do it :
- treat every stations as the leaf of a binary tree
- first slot (after successful transmission), all stations
can try to get the slot(under the root node). - If no conflict, fine.
- Else, in case of conflict, only nodes under a subtree get to try for the next one. (depth first search)
Adaptive Tree Walk Protocol fig (1.3)
Slot-0 : C*, E*, F*, H* (all nodes under node 0 can try which are going to send), conflict
Slot-1 : C* (all nodes under node 1 can try}, C sends
Slot-2 : E*, F*, H*(all nodes under node 2 can try}, conflict
Slot-3 : E*, F* (all nodes under node 5 can try to send), conflict
Slot-4 : E* (all nodes under E can try), E sends
Slot-5 : F* (all nodes under F can try), F sends
Slot-6 : H* (all nodes under node 6 can try to send), H sends.