Patent Application
20050195742
The following description relates to telecommunications in general and to asynchronous transfer mode (ATM) devices in particular.
Telecommunications networks typically use various multiplexing schemes to combine multiple data streams for transmission over a single physical transmission medium (also referred to here as a “link”). These schemes can be categorized into one of a number of different types. One scheme is time division multiplexing (TDM), which is commonly used in circuit-based applications.
In a TDM scheme, the single link has a data rate that is higher than the data rate of any of the multiple data streams that are to be combined. Typically, the link has a data rate that is at least N×f, where N is the number of data streams to be combined for transmission on the links and f is the data rate of each lower-speed, data stream. Each lower-speed data stream is assigned a fixed time slot during which only data from that input data stream can be transmitted on the link. This is the case regardless of whether or not that particular data stream actually has any data to be transmitted on the link during the time slot assigned to that data stream.
TDM multiplexing has been used in telecommunications networks for many years and serves as the basis for plesiosynchronous digital hierarchy (PDH) and synchronous digital hierarchy (SDH) technology. In both PDH and SDH technology, a higher speed link is divided into time slots or containers at fixed time intervals. These time slots are used for transporting telephone conversations, which are generally considered “synchronous” traffic; that is, the information is transmitted at regular intervals.
Other multiplexing schemes are more suitable for other types of data traffic. For example, data traffic tends to be “bursty.” That is, a given stream of data traffic is typically made up of bursts of data in between idle periods in which no data is transmitted. Combining multiple bursty data streams using a TDM scheme can cause bandwidth to be wasted. Bandwidth is wasted when no data is transmitted during a given time slot.
Other multiplexing schemes attempt to alleviate this issue with bursty traffic by transmitting data on the link only when there actually is data to be transferred for a give data stream. One such common multiplexing scheme is asynchronous transfer mode (ATM). ATM is commonly used in wide area telecommunications networks. In an ATM multiplexing scheme, data is partitioned into fixed size cells. Each cell has a fixed size header containing addressing, priority, error checking, and other routing information.
One common design goal of telecommunication service providers is to have a single network that is suited for transmission of both synchronous and bursty traffic. To this end, provisions were included in the definition of ATM to help make ATM suited for all types of network traffic. Within ATM, cells transported over the same link with the same header form a connection also referred to here as a “virtual circuit” or “VC”. When a connection is established and the data path is setup, a quality of service (QOS) is specified for that connection. The QOS determines the traffic characteristics of the connection. For example, synchronous traffic typically cannot tolerate varying delay and therefore is typically assigned a constant bit rate (CBR) QOS. A CBR QOS indicates that cells for that synchronous connection should be transferred through the switching and multiplexing equipment with a minimum delay and at a constant rate. Data or file transfer traffic (that is, bursty traffic) is typically assigned an unspecified bit rate (UBR) QOS. An UBR QOS indicates that cells for that connection should be transferred at a lower priority, without much concern for minimizing delay, and only when switching resources are available. Through the definition of these and other QOS levels, ATM equipment can handle different types of traffic.
ATM functions used to guarantee that QOS objectives are met include scheduling and shaping functions. Unlike in TDM schemes, the cells transmitted in ATM have no fixed time slot and can arrive at irregular intervals. The scheduling and shaping functions are traffic management functions that manage the transmission of cells from different connections. Cells from different connections are scheduled to insure that “higher priority” connections are transmitted first or at regular intervals. In addition, the combined rate of all the connections on a single link is shaped to guarantee that the combined rate does not exceed the physical data rate of the link.
Typically, these functions are implemented as scheduler blocks in one or more ATM processor semiconductor devices.
In the embodiment shown in
ATM multiplexer equipment such as digital subscriber loop access multiplexer (DSLAM) and broadband digital loop carrier (broadband DLC) equipment generally must support many physical layer termination devices, and hence need many schedulers. One approach is to use one scheduler block for each physical layer termination device. This, however, increases the cost of the equipment.
Alternatively, a single scheduler block can be used for multiple physical layer termination devices.
In this embodiment, the output of the scheduler block 200 is shaped to the combined rate of the multiple physical layer termination devices 208. For example, in the embodiment shown in
In one embodiment, a scheduler block includes a plurality of queues. Each queue is associated with at least one of a plurality of physical layer devices. Each queue stores packets intended for the physical layer device associated with that queue. Each packet has a priority level associated therewith and a weight indicative of the priority level of that packet. The scheduler block further includes an interface adapted to communicate with the plurality of physical layer devices. The scheduler block further includes a shaper, coupled to the plurality of queues and to the interface, that retrieves packets from the plurality of queues and forwards the packets to the interface. The order in which the packets are retrieved from the plurality of queues is based on the weight of each packet. For each physical layer device, the shaper shapes the packets retrieved from the queue associated with that physical layer device based on a data rate for that physical layer device.
In another embodiment, a method of forwarding scheduling packets, each packet intended for at least one of a plurality of physical layer devices and each packet being associated with a connection having a priority level, includes assigning each connection in a queue a weight related to the priority level of the queue. The method further includes, when a packet is received, enqueueing the received packet in the queue associated with the physical layer device to which the received packet is intended. The method further includes reading packets out of the plurality of queues in accordance with the weights assigned to the connections and shaping the packets read out of each queue to the data rate of the physical layer device associated with that queue. The method further includes outputting the packets.
In another embodiment, an access system includes a first unit including a first scheduler that performs traffic parameter based on scheduling for the access system. The access system further includes a plurality of second units coupled to the first unit. Each second unit includes a second scheduler block having a plurality of queues. Each queue is associated with at least one of a plurality of physical layer devices. Each queue stores packets intended for the physical layer device associated with that queue. Each packet has a priority level associated therewith and a weight indicative of the priority level of that packet. Each second unit further includes an interface adapted to communicate with the plurality of physical layer devices, and a shaper, coupled to the plurality of queues and to the interface, that retrieves packets from the plurality of queues and forwards the packets to the interface. The order in which the packets are retrieved from the plurality of queues is based on the weight of each packet. For each physical layer device, the shaper shapes the packets retrieved from the queue associated with that physical layer device based on a data rate for that physical layer device.
In another embodiment, a method of forwarding scheduling packets, each packet intended for at least one of a plurality of physical layer devices and each packet being associated with a connection having a priority level, includes, at a first unit, performing traffic parameter-based scheduling. The method further includes, at a second unit, assigning each connection in a queue a weight related to the priority level of the queue, and, when a packet is received, enqueueing the received packet in the queue associated with the physical layer device to which the received packet is intended. The method further includes, at the second unit, reading packets out of the plurality of queues in accordance with the weights assigned to the connections, shaping the packets read out of each queue to the data rate of the physical layer device associated with that queue, and outputting the packets. The second unit receives packets from the first unit in accordance with the traffic parameter-based scheduling.
The details of one or more embodiments of the claimed invention are set forth in the accompanying drawings and the description below. Other features and advantages will become apparent from the description, the drawings, and the claims.
Like reference numbers and designations in the various drawings indicate like elements.
Scheduler block 300 includes multiple queues 312. Each queue 312 is associated with a particular physical layer device 308. For example, in the embodiment shown in
The scheduler block 300 also includes a shaper 314 that is coupled to each of the queues 312. Shaper 314 retrieves cells from the queues 312 according to a weighted round robin scheme. Each connection has a “weight” assigned to that connection. The shaper 314 retrieves cells from the queues 312 in accordance with these weights in order to maintain the QOS levels of each connection.
In addition, the shaper 314, for each queue 312 that the shaper 314 reads cells from, also shapes the cells read from that queue 312 to the data rate of the physical layer device 308 associated with that queue 312. In one implementation, the cells read out of each queue are shaped according to a weighted round robin (WRR) algorithm in accordance with standards promulgated by the ATM Forum.
In one embodiment, the scheduler block 300 is implemented as one or more application-specific integrated circuits (ASIC). Other embodiments are implemented in other ways, for example, using a field programmable gate array (FPGA), discrete components, programmable processors and the like.
Method 400 includes assigning each connection in a queue a weight related to the QOS of that connection (or other priority) (block 402). Weights are assigned to maintain the relative priority of the connections. For example, in one implementation, a connection having a CBR QOS is assigned a weight that is twice as large as the weight assigned to a connection having a UBR QOS.
When a cell is received (determined in block 404), the received cell is enqueued in the queue associated with the physical layer termination device for which the cell is intended (block 406). In one embodiment, another device (for example, a central unit described below in connection with
When a cell is in a queue (which is checked in block 408), cells are read out of the queues in accordance with the assigned weights (block 410). For example, in one embodiment, the cells are read out of the queues in accordance with a weighted round robin scheme as described above. The cells read out of each queue are shaped to the data rate of the physical layer termination device 308 associated with that queue (block 412). In one implementation, the cells read out each queue are shaped according to a WRR in accordance with standards promulgated by the ATM Forum. In one embodiment, the operations of block 408, block 410, and block 412 are performed in parallel with the operations of block 402, block 404, and block 406, for example, by using dual-ported random access memory (RAM).
With embodiments of the scheduler block 300 and method 400, data flows intended for different physical layer termination devices need not affect each other, and limited bursts will be shaped so that the physical layer termination device are better able to handle the burst without loss. Each of the queues 312 is associated with, and serves, a separate physical layer termination device. In other words, in a network device in which thirty two scheduler blocks 300 are used, where each block has five queues data streams intended for one hundred sixty physical layer termination devices (in other words, thirty two times five) are supported. Such an approach can reduce the costs associated with providing such scheduling and shaping functionality.
Access system 500 includes a central unit 502 and multiple remote units 504. The central unit 502 includes a network interface (for example, a network-node interface (NNI) or user-network interface (UNI)) 506 that couples the central unit and the access system 500 to an ATM network (not shown). For example, in one embodiment, a SONET ring is coupled to the network interface 506. Each of the remote units 504 includes multiple user interfaces (UNIs) 508. Each user interface 508 couples a remote unit 504 and the access system 500 to one of multiple edge devices (not shown). For example, in one embodiment, an asynchronous digital subscriber line (ADSL) modem or a broadband integrated services digital network (B-ISDN) modem is coupled to at least one of the user network interfaces 508. The central unit 502 and each of the remote units 504 includes an input inter-unit interface 510 and an output inter-unit interface 512 that are used to couple each unit to two other units in the ring. For each unit in the access system 500, cells are received from another unit over the input inter-unit interface 510 and cells are transmitted to another unit over the output inter-unit interface 512. In one embodiment, the network interface 506 is a point-to-point SONET UNI/NNI, the inter-unit interfaces 510 and 512 are SDH/ATM interfaces, and the user network interfaces 508 include ADSL and/or B-ISDN interfaces.
The central unit 502 includes at least one central scheduler block 514 that perform all traffic parameter-based scheduling for the access system 500. Cells received from the network interface 506 and the input inter-unit interface 510 are scheduled, shaped, and ultimately output by the central unit 502 on the network interface 506 and the output inter-unit interface 512 as appropriate. The central scheduler block 514 performs full traffic parameter-based scheduling for the cells the central scheduler block 514 processes.
Each remote unit 504 includes at least one remote scheduler block 516. Remote scheduler block 516 is an embodiment of the scheduler block 300 of
In one embodiment, each remote unit 504 includes an ATM interface card and one or more ADSL or B-ISDN line interface cards. In such an embodiment, the ATM scheduling card includes the remote scheduler block 516, the input inter-unit interface 510, the output inter-unit interface 512, and the ATM switching device 518. Alternatively, the inter-unit interfaces 510 and 512 can be on different cards. Each line interface card in such an embodiment includes the user interfaces 508. In such an embodiment, the ATM interface card and the one or more ADSL or B-ISDN line interface cards are coupled to each other over a backplane.
The traffic management parameters are processed in the central unit 502, while the remote units 504 need not be designed to carry out such processing of traffic management parameters. The remote units 504, therefore, can be relatively simple and have static configurations that assign appropriate weights (for example, low or high weights) to preserve the scheduling done at the central unit 502. Such an approach reduces the need to manage many configuration tables in the central unit 502 and the remote units 504. Instead, a single configuration table is managed at the central unit 502, while the remote unit's simply assign the appropriate weight based on the QOS type that is identified by the VPI or VCI in the cell header.
Method 600 includes, at a first unit, performing traffic parameter-based scheduling (block 602). In an embodiment implemented using an embodiment of access system 500, traffic parameter-based scheduling is performed in central unit 502 by central scheduler block 514. For example, in an implementation of such an embodiment, during setup each connection is assigned an ATM layer service category which best complies with the connection's required QOS and the expected behavior of the connection. The service category dictates which connection traffic parameters are specified and the QOS objectives. Once established, the central unit 502 monitors the connection and schedulers cells to attempt to ensure that the connection conforms to the traffic parameters.
Method 600 also includes operations executed at a second unit. Method 600 includes, at the second unit, assigning each connection in a queue a weight related to the QOS of that connection (or other priority) (block 604). Weights are assigned to maintain the relative priority of the connections. For example, in one implementation, a connection having a CBR QOS is assigned a weight that is twice as large as the weight assigned to a connection having a UBR QOS.
At the second unit, when a cell is received (determined in block 606), the received cell is enqueued in the queue associated with the physical layer termination device for which the cell is intended (block 608). The first unit performs traffic-parameter-based scheduling and provides the packets (for example, ATM cells) cells to the second unit in accordance with the results of that scheduling.
When a cell is in a queue at the second unit (which is checked in block 610), cells are read out of the queues in accordance with the assigned weights (block 612). For example, in one embodiment, the cells are read out of the queues in accordance with a weighted round robin scheme as described above. The cells read out of each queue are shaped to the data rate of the physical layer termination device associated with that queue (block 614). In one implementation, the cells read out each queue are shaped according to a WRR in accordance with standards promulgated by the ATM Forum. In one embodiment, the operations of block 610, block 612, and block 614 are performed at the second unit in parallel with the operations of block 604, block 606, and block 608, for example, by using dual-ported random access memory (RAM).
The methods and techniques described here may be implemented in digital electronic circuitry, or with a programmable processor (for example, a special-purpose processor or a general-purpose processor such as a computer) firmware, software, or in combinations of them. Apparatus embodying these techniques may include appropriate input and output devices, a programmable processor, and a storage medium tangibly embodying program instructions for execution by the programmable processor. A process embodying these techniques may be performed by a programmable processor executing a program of instructions to perform desired functions by operating on input data and generating appropriate output. The techniques may advantageously be implemented in one or more programs that are executable on a programmable system including at least one programmable processor coupled to receive data and instructions from, and to transmit data and instructions to, a data storage system, at least one input device, and at least one output device. Generally, a processor will receive instructions and data from a read-only memory and/or a random access memory. Storage devices suitable for tangibly embodying computer program instructions and data include all forms of non-volatile memory, including by way of example semiconductor memory devices, such as EPROM, EEPROM, and flash memory devices; magnetic disks such as internal hard disks and removable disks; magneto-optical disks; and DVD disks. Any of the foregoing may be supplemented by, or incorporated in, specially-designed application-specific integrated circuits (ASICs).
A number of embodiments of the invention defined by the following claims have been described. Nevertheless, it will be understood that various modifications to the described embodiments may be made without departing from the spirit and scope of the claimed invention. For example, although embodiments involving scheduling and shaping ATM cells are described above, it is to be understood that other types of packets are used in other embodiments. Accordingly, other embodiments are within the scope of the following claims.
