Patent Grant
7583663
A. Field of the Invention
The present invention relates generally to packet processing and, more particularly, to converting a first data path that carries up to P packets per processing cycle to a second data path that carries Q packets per processing cycle, where Q<P.
B. Description of Related Art
Packet processing systems, including any type of router, server or host that communicates using a packet-switching access mechanism, conventionally receive and process multiple packets in a single system cycle. A packet processing system may include a wide data path for receiving multiple packets in parallel during a single system cycle. Processing the packets at the rate they are received may require multiple instances of processing logic operating in parallel. If the packet processing system receive data path is N bytes wide, and the minimum packet that must be processed is M bytes, then P=N/M instances of the processing logic may be required to process all packets in a given system cycle. For example, conventional Cyclical Redundancy Checking (CRC) may be performed to determine packet data errors.
Multiple instances of processing logic in the packet processing system, however, may have many drawbacks, such as increased power demands and space requirements in the system. In Application Specific Integrated Circuits (ASICs), for example, multiple instances of processing logic utilize valuable area of the ASIC. Multiple elements operating in parallel also increase timing complexity in the system.
To decrease space and power requirements in the packet processing system, it would, thus, be desirable to reduce the instances of the logic required to process multiple packets received during a single system cycle. For example, reduction of the processing logic to, for example, a single instance would significantly reduce space and power requirements. Therefore, to enable the use of a single instance of packet processing logic, there exists a need for systems and methods that can convert a data path carrying P packets per cycle to a data path that carries only Q packets per cycle, such as Q=1 packet per cycle.
Consistent with the principles of the invention disclosed and claimed herein, these and other needs are addressed by queuing up to P packets per cycle received on a first data path and outputting Q packets per cycle on a second data path, where Q<P. Thus, in one embodiment, the packet processing system may use only a single instance of processing logic, such as, for example, CRC logic to process the P packets.
According to one aspect, a method may include storing P data units, each comprising multiple bytes, in a queue. The method may further include shifting, byte by byte, Q data units from the queue during a first system clock cycle, wherein Q<P and sending, during the first system clock cycle, the Q data units to a processing device configured to process a maximum of Q data units per system clock cycle.
According to another aspect, a system may include a queue that stores P data units, each data unit including multiple bytes. The system may further include a control unit that shifts, byte by byte, Q data units from the queue during a first system clock cycle, wherein Q<P, and sends, during the first system clock cycle, the Q data units to a processing device configured to process a maximum of Q data units per system clock cycle.
According to an additional aspect, a system may include a queue configured to receive and store first P packets during a first system clock cycle and a processing device configured to process a maximum of Q packets per system clock cycle. The system may further include a control unit configured to selectively retrieve first Q packets of the first P packets from the queue during the first system cycle, wherein Q<P, and send the Q packets to the processing device for processing during a single system clock cycle.
According to yet another aspect, a system may include a clock device configured to produce clock signals having a system clock cycle, a data path configured to carry a maximum of P data units per system clock cycle and a data processing device configured to process a maximum of Q data units per system clock cycle. The system may further include one or more devices connected to the data path and the data processing device and configured to receive P data units from the data path during a first system clock cycle, and supply, during the first system clock cycle, first Q data units of the P data units to the data processing device for processing, wherein Q<P.
According to a further aspect, a method may include receiving P data units during a first system clock cycle, where each of the P data units includes multiple bytes, and storing the P data units in a queue. The method may further include selectively retrieving bytes of the P data units from the queue until Q data units have been retrieved, where Q<P, and sending, during the first system clock cycle, the Q data units to a data processing device capable of handling a maximum of Q data units per clock cycle.
According to another aspect, a system may include a queue configured to receive P data units during a first system clock cycle, where each of the P data units includes multiple bytes, and store the P data units. The system may further include a control unit configured to selectively retrieve bytes of the P data units from the queue until Q data units have been retrieved, where Q<P, and send, during the first system clock cycle, the Q data units to a data processing device capable of handling a maximum of Q data units per clock cycle.
The accompanying drawings, which are incorporated in and constitute apart of this specification, illustrates an embodiment of the invention and, together with the description, explain the invention. In the drawings,
The following detailed description of the invention refers to the accompanying drawings. The same reference numbers in different drawings may identify the same or similar elements. Also, the following detailed description does not limit the invention. Instead, the scope of the invention is defined by the appended claims and equivalents.
Systems and methods consistent with the present invention provide mechanisms that can queue up to P packets per cycle received on a first data path and output Q packets per cycle on a second data path, wherein Q<P. In one embodiment, Q=1, thus, enabling the use of a single instance of packet processing logic to process the received P packets.
Packet processing system 100 may include a data path 105, a FIFO queue 110, a processing unit 115, a data path 120, a processing device(s) 125, and a clock 130. Data path 105 may be N data units, such as N bytes, in width and may include conventional mechanisms for receiving up to P packets per processing cycle and for transmitting the received packets to FIFO queue 110. FIFO queue 110 can receive and store packets received from data path 105 and may include one or more memory devices (not shown).
Processing unit 115 may include a processor that contains instructions for processing data received from FIFO queue 110. Processing unit 115 may alternatively include logic circuitry that performs the equivalent functions of a processor. Processing unit 115 also may include flow control feedback to FIFE queue 110.
Processing device(s) 125 may include a device that processes up to Q incoming packets per processing cycle. For example, processing device(s) 125 may include a Cyclical Redundancy Check (CRC) unit that performs conventional CRC data error checks upon packets received via data path 105. Processing device(s) 125 may also include multiple devices that can each process a maximum of Q packets per cycle, where each of the multiple devices may perform different functions. For example, processing device(s) 125 may include a checksum unit, a CRC unit and a FIFE unit, with each unit processing a maximum of Q packets/cycle, such as 1 packet/cycle.
Clock 130 includes conventional circuitry for supplying clock signals to the components of packet processing system 100. Clock 130 may, for example, supply a clock signal to FIFO queue 110, processing unit 115, and processing device(s) 125.
Turning to
If byte c of the shift register indicates that it is the last byte of a packet, byte C may be retrieved from the shift register (step 435). A packet may then be sent including the retrieved bytes on a 1 packet/cycle data path (step 440) at, or before, completion of the current processing cycle defined by the clock signal (step 445).
Consistent with the principles of the present invention, a data path carrying more than one packet per system cycle may be converted to a data path that carries only Q packets per system cycle, such as 1 packet/cycle. Thus, packet processing may be performed using only a single instance of processing logic, such as CRC logic, to process multiple packets received during a single system cycle.
The foregoing description of preferred embodiments of the present invention provides illustration and description, but is not intended to be exhaustive or to limit the invention to the precise form disclosed. Modifications and variations are possible in light of the above teachings or may be acquired from practice of the invention. For example, while series of steps have been described with regard to
The scope of the invention is defined by the claims and their equivalents.
The present application is a continuation of U.S. application Ser. No. 09/895,218, filed Jul. 2, 2001, and issued as U.S. Pat. No. 7,103,038, the disclosure of which is incorporated by reference herein in its entirety.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4556850 | McBrien | Dec 1985 | A |
| 4907225 | Gulick et al. | Mar 1990 | A |
| 5220325 | Ackland et al. | Jun 1993 | A |
| 5365519 | Kozaki et al. | Nov 1994 | A |
| 5381409 | Folkert de Vries | Jan 1995 | A |
| 5768630 | Kim | Jun 1998 | A |
| 5790786 | Wakeman et al. | Aug 1998 | A |
| 6385671 | Hunsaker et al. | May 2002 | B1 |
| 6421348 | Gaudet et al. | Jul 2002 | B1 |
| 6509851 | Clark et al. | Jan 2003 | B1 |
| 6650638 | Walker et al. | Nov 2003 | B1 |
| 7103038 | Gaudet | Sep 2006 | B1 |
| 20010012290 | Kanazashi | Aug 2001 | A1 |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 09895218 | Jul 2001 | US |
| Child | 11458809 | US |
