* N1 needs 300 msec to transmit 1 MB of data
* N2 needs 500 msec to transmit 1.2 MB of data
* N3 needs 200 msec to transmit 2 MB of data
* The shortest deadline is for N3 which is 200 msec
* According to IEEE 802.4, TTRT should be less than the shortest deadline/2 = 200/2 = 100 msec
* With TTRT of 100 msec, synchronous bandwidths are:
- N1: 1 MB/100 msec = 10 Mbps
- N2: 1.2 MB/100 msec = 12 Mbps
- N3: 2 MB/100 msec = 20 Mb
* N1 needs 300 msec to transmit 1 MB of data
* N2 needs 500 msec to transmit 1.2 MB of data
* N3 needs 200 msec to transmit 2 MB of data
* The shortest deadline is for N3 which is 200 msec
* According to IEEE 802.4, TTRT should be less than the shortest deadline/2 = 200/2 = 100 msec
* With TTRT of 100 msec, synchronous bandwidths are:
- N1: 1 MB/100 msec = 10 Mbps
- N2: 1.2 MB/100 msec = 12 Mbps
- N3: 2 MB/100 msec = 20 Mb
* N1 needs 300 msec to transmit 1 MB of data
* N2 needs 500 msec to transmit 1.2 MB of data
* N3 needs 200 msec to transmit 2 MB of data
* The shortest deadline is for N3 which is 200 msec
* According to IEEE 802.4, TTRT should be less than the shortest deadline/2 = 200/2 = 100 msec
* With TTRT of 100 msec, synchronous bandwidths are:
- N1: 1 MB/100 msec = 10 Mbps
- N2: 1.2 MB/100 msec = 12 Mbps
- N3: 2 MB/100 msec = 20 Mb
* N1 needs 300 msec to transmit 1 MB of data
* N2 needs 500 msec to transmit 1.2 MB of data
* N3 needs 200 msec to transmit 2 MB of data
* The shortest deadline is for N3 which is 200 msec
* According to IEEE 802.4, TTRT should be less than the shortest deadline/2 = 200/2 = 100 msec
* With TTRT of 100 msec, synchronous bandwidths are:
- N1: 1 MB/100 msec = 10 Mbps
- N2: 1.2 MB/100 msec = 12 Mbps
- N3: 2 MB/100 msec = 20 Mb
• In Ethernet, under high load situations: • Increased delays, and drop in throughput.
• Number of collisions increase rapidly
• Leads to futile retransmissions.
• limited use of Ethernet networks in real-time applications.
Hard Real-Time Communication in LAN
• Hard real-time applications often involve transmission of (constant bit
rate) CBR traffic.
• Network utilization is kept low
• Predictability of delays is more important than utilization aspects.
• Soft and non real-time traffic are allowed to be transmitted when
hard real-time messages are not transmitted. Real Time communication in LAN • Three main categories of protocols:
• Global priority protocols
• Bounded access scheduling • Calendar-based protocols Global Priority Protocols • Each message is assigned a priority value.
• The MAC layer protocol tries to ensure that at any instant of time, only the highest priority message is being served.
• RMA or EDF can then be used to schedule messages.
• However several issues make implementing RMA or EDF-based scheduling inefficient. Task versus Packet Scheduling • After the transmission of a packet starts, it cannot be stopped halfway.
• Unlike tasks, packet transmissions cannot be meaningfully
preempted.
• A global protocol: • Cannot instantaneously determine the highest priority message Calendar- Based Protocol
• A calendar is maintained:
• Indicates the time period during which a node is allowed to transmit.
• A copy of the calendar is maintained at every node.
• It contains information about the access times reserved for
guaranteed messages from all the nodes. Calendar based Protocol. • When a message with no reserved slot arrives,
• The node determines a free sIot by consulting the local calendar.
• It then broadcasts a control message.
• Other nodes update their local calendars accordingly.
Calendar Based Protocol • A dynamic reservation technique.
• Calendar-based protocol: • Efficient for periodic messages.
• It is very difficult to handle aperiodic and sporadic messages.
Bounded Access Protocols • The access time of every node to the channel is bounded.
• This limits the time for a packet to be transmitted.
• Individual nodes use a local scheduling algorithm:
• determine the order in which queued up packets are taken up for
transmission. Count down protocol • It is a type of global priority-based protocol. • The time line is divided into fixed size intervals called slots.
• At the start of each slot:
• Priority arbitration is carried out to
determine the highest priority message.
• The node having the highest priority
message is allowed to transmit. Count down protocol • Choosing appropriate slot size: critical for efficient working of the protocol. IEEE 802.5 • Priority based token ring protocol.
• The header of the token contains two fields:
• Reservation and Mode fields.
• The token alternates between two modes:
• Reservation mode • Free mode IEEE 802.5
• IEEE 802.5 supports assigning priorities to messages.
• Priority is stored in reservation field.
• Messages are split into frames.
• As token circulates, reservation field is checked by each node.
• A node having higher priority message records its priority in the
header. IEEE 802.5 • The sender node checks the reservation status of the token when the token returns.
• If some other higher priority node has registered, it frees the token.
• After one complete circulation:
• The high priority node that made the reservation captures the token and starts transmitting. IEEE 802.4 • Important points:
• It can be used in both in token ring and token bus networks.
• It is also known as the timed token protocol.
• A node can transmit only when it holds the token.
• The time for which a node can transmit is bounded.
• HeIp in providing real-time guarantees. IEEE 802.4 • Target Token Rotation Time (TTRT): • Expected time between two consecutive visits of a token to a node.
• This is an important design parameter. Initialized during network
setup
• Real-time messages are assumed to be periodic
• Also called synchronous messages. • While non real-time messages are called Asynchronous messages. IEEE 802.4 • Each individual node is allocated a portion of the synchronous bandwidth.
• When a node receives the token;
• It transmits real time messages for a time less than the synchronous bandwidth.
• After completing transmission of the synchronous traffic.
• The node can transmit asynchronous messages. IEEE 802.4 • Synchronous bandwidth of node Ni be Hi and θ is the propagation time, SN is set of all nodes in the network Then : IEEE 802.4 • Worst case time between two successive visits of token to node is • 2*TTRT
• Assume node Ni has a message with the shortest deadline Δ,
• TTRT should be set to a value lower than Δ/2, during network utilization • A network designed using IEEE802.4 protocol has three nodes: • N1 needs to transmit 1 MB of data every 300 msec. • N2 needs to transmit 1.2 MB of data every 500 msec. • N3 needs to transmit 2 MB of data every 200 msec.
• Select suitable TTRT?
• Determine the synchronous bandwidth of the nodes or token holding time.
