ATM Tutorial: Network Layer

I. Network Layer in ATM
The ATM layer processes the cells moving from the source end to the target end. The ATM switch does contain routing algorithms and protocols, and it also handles global addressing issues. Therefore, the ATM layer functions the same as the network layer. The ATM layer does not guarantee reliability, but it is not required for a network layer protocol.
Because the ATM layer has the functions of the network layer rather than the functions of the data link layer, and the ATM layer is similar to the existing network layer, we will continue to discuss the ATM layer protocol in this chapter.
The only problem is that the ATM layer does not have the characteristics of the data link layer protocol: A single station protocol used between machines at both ends of the wire, just like Protocol 1 to Protocol 6 in Chapter 3rd. The ATM layer has the functions of network layer protocols: End-to-End virtual circuit connection, switching, and route selection.
For connection-oriented protocols, the ATM layer is unusual because it does not provide any validation. However, the ATM layer still provides powerful protection: the cells sent along a virtual circuit will never lose the order. If blocking occurs, the ATM subnet is allowed to discard cells, but under no circumstances can it re-sort the cells transmitted in a separate virtual circuit. However, cells transmitted in different virtual circuits are not guaranteed in order.
Ii. Cell format
In the ATM layer, two interfaces are very important, namely user-network interface (UNIuser-network interface) and network-network interface (NNInetwork-network interface ). The former defines the border between the host and the ATM network in many cases between the customer and the carrier), and the latter applies to the routers in the sense of two ATM switches. Cell transmission takes precedence over the leftmost bytes, and within a byte, the leftmost bits take precedence.
3. Establish a connection
Technically speaking, connection establishment is not part of the ATM layer, but is handled by a highly complex ITU protocol called Q.2931 (stiller, 1995) used by the control platform. However, logically, the network layer connection is established at the network layer, and similar network layer protocols are established here, so we will discuss it here.
Messages used for connection establishment and connection release
The meaning of a message sent by the host is the meaning of the message sent by the network.
Set up a virtual circuit to enter the call
CALL PROCEEDING
Connect I accept incoming call accept your call request
Connect ack thank you for accepting. Thank you for calling.
RELEASE, please terminate the call. The other end is bad enough.
Release complete confirmation of RELEASE
The ATM network allows the establishment of multi-point broadcast channels. A multi-point Broadcast Channel has one sender and more than one receiver. They are established through the following method: use the common method to establish a connection between the source end and the target end, then, send the add party message to connect the second destination to the virtual circuit returned by the previous call. Then, you can send other ADD parties to increase the number of the destination.
ATM has three address formats. 1st bytes indicates which of the three address formats the address belongs. The 1st is 20 bytes long and is based on the OSI address format. 2nd bytes and 3rd bytes indicate the country, 4th bytes indicate the address-based format, others include 3 bytes specify the permission, 2 bytes specify the domain), 1 byte specify the region, there are also 6-byte addresses and other information items. In the 2nd address formats, 2nd and 3rd bytes specify an international organization rather than a country; the remaining part and format of the address are the same as those of 1st. The other is that the old ISDN phone number CCITT E.164 in 15 decimal digits is used as the address format.
4. Route Selection and switching
When a virtual circuit is established, the SETUP message goes from the source end to the destination end of the network. The routing selection algorithm determines the path of the message, which determines the path of the virtual circuit. No specific routing algorithm is specified in the ATM standard. Therefore, you can select one of the routing algorithms discussed in the previous sections of this chapter, or use another algorithm.
Most of the work of a vswitch is spent on how to get the output line from the virtual circuit information in a cell. Except for the last station segment in each direction, routes are carried out in the VPI field, rather than in the VCI field. In the last station segment, cells are transmitted between the switch and the host. Only virtual channels are used between two vswitches.
In Lan, things are much simpler. A simple virtual path can be used by all virtual circuits.
V. service type
Constant bit rate (CBRconstant bit rate) is mainly used to imitate copper wires or optical fibers. No error verification, no traffic control, and no other processing. This category makes a smooth transition between the current telephone system and the future B-ISDN system, because of the voice-level PCM channel, t1 circuit and other telephone systems all use synchronous data transmission at a constant rate.
Variable bit rate (VBRvariable bit rate) is divided into two sub-groups, which are set up for real-time transmission and non-real-time transmission. RT-VBR is mainly used to describe services that have variable data streams and require strict real-time, such as interactive compressed videos such as video conferencing ). NRT-VBR is used in communication scenarios that are mainly scheduled to send messages. In such cases, a certain number of latencies and their changes can be tolerated by applications, such as emails.
Available bit rate ABRavailable bit rate) the term is designed for sudden information transmission that is generally known to the bandwidth range. The only type of service that the network will provide speed feedback to the sender. When network congestion occurs, the sender is required to reduce the transmission rate. Assuming that the sender complies with these requests, the loss of cells using the ABR communication will be very low. The operational ABR is a bit like a mobile passenger waiting for opportunities: if there is a free seat space), the mobile passenger will be delivered to the free seat without delay; if there is not enough capacity, they have to wait unless some minimum bandwidth is available ).
The bit rate UBRunspecified bit rate is not specified, and there is no feedback on congestion. This type is suitable for sending IP datagram. In the event of congestion, UBR cells will also be discarded, but will not send feedback to the sender or send the sender's expectation of slowing down.
Features of various ATM service types
Service Characteristics CBR RT-VBR NRT-VBR ABR UBR
Bandwidth guarantee is optional
Applicable to real-time communication.
Applicable to burst communication.
No, no, no
Vi. Service Quality
Service quality is an important topic in the ATM network, because the ATM network is used for real-time transmission, such as audio and video. When a virtual circuit is established, the transport layer is typically a process in the host, "customer") and the ATM network layer, for example, a network operator, that is, a "carrier provider ") must comply with a service definition agreement.
The first part of the protocol is the traffic descriptor ). It describes the load to be provided. The second part of the Agreement specifies the Quality of Service requested by the customer and agreed by the communication provider. Both loads and services must be described in a measurable quantity, so that the conventions can be objectively determined.
In order to make specific traffic agreements possible, the ATM standard defines a series of quality of service (Qos), and customers and communication providers can negotiate the value of these parameters. For each service quality parameter, the worst case value is specified, requiring the communication provider to reach or exceed this value. In some cases, a parameter is a minimum value, while in other cases it is a maximum value. Here, the service quality is specified separately in every direction. Some of the important columns are in the following table, but they are not applicable to all service types.
Some service quality parameters
Parameter Acronyms
Peak cell rate the maximum rate at which PCR cells are sent
Continuous cell rate (SCR) Long-time average cell transmission rate
Minimum cell rate minimum acceptable cell transmission rate MCR
Maximum number of acceptable cell jitter when the cell delay changes
Cell Loss Rate CLR cell loss or too late submission Ratio
The Intermediate Value and maximum value of the delay time when the CTD cell is submitted)
Cell delay variation CDV cell submission time variation margin
Number of error cells submitted by CER
Serious Error cell block rate: proportion of cells with SECBR errors
Cell error destination rate the ratio of CMR cells submitted to the wrong destination
VII. Traffic Shaping and Control
The mechanism for using and enhancing the service quality parameters is partly based on) a specific algorithm, that is, the general cell rate algorithm (generic cell rate algorithm ). It works by checking every cell to see if it complies with the parameters of the virtual circuit.
GCRA has two parameters that specify the maximum allowed arrival rate PCR) and the tolerable arrival time variation CDVT ). The reciprocal of PCR, T = 1/PCR is the minimum cell arrival interval value.
The routing a algorithm is called virtual scheduling algorithm. However, from another perspective, it is equivalent to a bucket leakage algorithm. Think of an agreed cell as a fluid that is poured into a T unit of a drain bucket. This bucket leaks liquid at a rate of 1 unit/us, so it is empty after Tus. If the cell arrives at exactly the speed of 1 cell/Tus, then each arriving cell will find that the bucket is just empty, and the cell will reload the T unit of liquid in the bucket. Therefore, when a cell arrives, the water level of the liquid increases to T, and then decreases linearly until zero.
When a cell arrives Lus in advance, the bucket should overflow. For a given T, if we set L very small, the bucket capacity will be difficult to exceed T, so all cells must be sent at a very standard interval. However, if we increase the value of L so that it is much larger than T, the bucket will hold a lot of cells because T + L> T. This means that the sender can send burst data one by one at the peak rate, and they can still be correctly received.
Generally, distinct A is specified by specifying the T and L parameters. T is the reciprocal of PCR; L is CDVT. Latency A is also used to ensure that the average cell transmission rate does not exceed SCR in any period of time.
In addition to providing a rule to see which cell is compliant with the agreement and which one is not compliant with the agreement, limit A is also used for communication shaping to eliminate some sudden transmission. The smaller the CDVT, the better the smoothness, but the higher the chance of discarding cells due to noncompliance. In some implementations, combine the drain a bucket with a token bucket to provide further smoothing.
VIII. Congestion Control
The ATM network must not only deal with long-term congestion caused by traffic greater than the system's processing capability, but also deal with short-term congestion caused by sudden transmission in communications. As a result, people use several different strategies. The most important of them can be divided into three categories:
1. License Control
Many ATM networks have real-time sources that generate data at a fixed rate. It is impossible to tell this type of communication source to slow down the sending rate. Imagine a new type of digital phone with a red light. When a notification congestion occurs, the red light will be on, and the speaker will be asked to speed down by 25% ).
Therefore, the ATM network puts congestion prevention first. However, there is no Dynamic Congestion Control for CBR, VBR, and UBR traffic, so preventing congestion here is much stronger than restoring it after congestion occurs. A major tool to prevent congestion is license control. When a host needs a new virtual circuit, it must describe the communication and service to be provided, and the network will check whether it is possible, handle the connection without affecting existing connections. You may need to check multiple possible lines to find which one can do the job.
2. Resource Reservation
Skills that are closely related to license control are used to reserve resources in advance, which is typically performed when a call is established. Because the traffic descriptor provides the cell transmission peak rate, the Network may reserve enough bandwidth along the path to process the peak rate.
3. Speed-based Congestion Control
In the communication between CBR and VBR, because of the inherent real-time and semi-Real-Time Characteristics of the information source, even in the case of congestion, it is generally impossible for the sender to slow down the transmission rate. No one worries about the VBR Service. If there are too many cells, discard them.
In An ABR communication, it is possible that the network notifies one or more senders and requests them to temporarily slow down the sending rate until the network recovers.
How to detect, notify, and control the congestion in the ABR communication is a hot topic in the development of the ATM standard. The problems are concentrated in the following two aspects: first, credit-based solutions, one is speed-based solutions.
Switch manufacturers oppose credit-based solutions. They don't want to perform all calculations to remember these credits, nor do they want to provide a lot of buffers in advance, and think that the amount of overhead required is too large. Therefore, a speed-based congestion control system is adopted. The basic model is that each sender transmits a special resource management RM (resource management) cell after k-cell data. The transmission path of this cell is the same as that of k cell, but it is specially processed by the switch. When an RM cell reaches the receiving end, it is detected, modified, and sent back to the sending end. In addition, two other congestion control devices are provided. The first is that the superload switch can generate RM cells spontaneously and send them back to the sender. The second is that the superload switch can set the intermediate PTI value for the cell data transmitted from the sending end to the receiving end. Of course none of the two methods is completely reliable.