Merge systems and methods for transmit system interfaces

• US 8,621,100 B1
• Filed: 02/27/2009
• Issued: 12/31/2013
• Est. Priority Date: 07/18/2002
• Status: Expired due to Fees

• Current Assignee: Juniper Networks Incorporated
• Original Assignee: Juniper Networks Incorporated
• Inventors: Yeluri, Sharada, Clark, Kevin, Ilislamloo, Shahriar, Lau, Chung
• Primary Examiner(s): Daftuar, Saket K
• Application Number:
  US12/394,620
• Time in Patent Office: 1,768 Days
• Field of Search: 370412-421, 370/428, 370/252, 370/465, 370/503, 370/229, 370/235, 370/466, 370/468, 370/470, 370/472, 370/476, 370/477, 370/505, 370/517, 370/907, 710/52, 710 56- 57, 710 71- 72, 710/307, 710/310, 710/29, 710/30, 710/33, 710/34, 710/35, 710/308, 709/230, 709/231, 709/232, 709/233, 709/234, 709/235, 709/236
• US Class Current: 709/234
• CPC Class Codes: H04L 45/00 Routing or path finding of packets in data switching networks routing or path finding in wireless networks H04W40/00
• CPC Class Codes: H04L 45/60 Router architectures
• CPC Class Codes: H04L 47/10 Flow control; Congestion control
• CPC Class Codes: H04L 49/254 Centralised controller, i.e. arbitration or scheduling
• CPC Class Codes: H04L 49/3072 Packet splitting
• CPC Class Codes: H04L 49/90 Buffering arrangements
• CPC Class Codes: Y10S 370/907 Synchronous optical network, SONET

Abstract

A system improves bandwidth used by a data stream. The system receives data from the data stream and partitions the data into bursts. At least one of the bursts includes one or more idles. The system selectively removes the idles from the at least one burst and transmits the bursts, including the at least one burst.

Claims

1. A transmit interface comprising:

a plurality of schedulers, implemented at least partially in hardware, to;

partition data, received via a plurality of data streams, into bursts;

a buffer, associated with the plurality of schedulers and a data stream of the plurality of data streams, to;

determine that at least one burst, associated with the data stream, includes one or more idles,

compare a current capacity of the buffer to a buffer threshold value, and

determine, based on the comparing, that the one or more idles are to be removed from the data stream; and

merge logic, implemented at least partially in hardware, to;

selectively pack, based on the determination that the one or more idles are to be removed, the at least with data from another data stream, of the plurality of data streams, to remove the one or more idles from the at least one burst.

2. The transmit interface of claim 1, where the merge logic is further to:

multiplex the bursts, including the at least one burst, to create multiplexed streams, and

where the transmit interface further comprises;

a transmitter to transmit the multiplexed streams.

3. The transmit interface of claim 1, where merge logic includes:

one or more multiplexers to;

receive the bursts, including the at least one burst, from the plurality of schedulers, and

selectively pack the at least one burst, and an asynchronous buffer to;

store the bursts.

4. The transmit interface of claim 3, where the merge logic is further to:

write the at least one burst to the asynchronous buffer at a first frequency, and

read the at least one burst from the asynchronous buffer at a second frequency,

the second frequency being different from the first frequency.

5. The transmit interface of claim 3, where the asynchronous buffer includes:

a current pointer to;

identify next data to be read from the asynchronous buffer, and fill logic to;

compare the current pointer to the buffer threshold value, and generate a control signal to instruct the one or more multiplexers to perform no data packing when the current pointer is less than the buffer threshold value.

6. The transmit interface of claim 5, where, when the one or more multiplexers perform no data packing, the one or more multiplexers are to:

store data from the bursts in the asynchronous buffer without removing idles present in the bursts from the bursts.

7. The transmit interface of claim 3, where the one or more multiplexers are to selectively pack the at least one of the bursts to remove the one or more idles from the at least one of the bursts, and

where, based on selectively packing the at least one of the bursts, the one or more multiplexers are to:

store data associated with the at least one of the bursts in the asynchronous buffer with no idles.

8. A method, performed by one or more devices, the method comprising:

partitioning, by the one or more devices, data, received from a plurality of data streams, into bursts;

determining, by the one or more devices and based on the partitioning, that at least one burst, associated with a data stream, of the plurality of data stream, includes one or more idles;

comparing, by the one or more devices, a current capacity of a buffer associated with the data stream to a buffer threshold value; and

packing, by the one or more devices and when the current capacity of the buffer is not less than the buffer threshold value, the at least one burst with data from another data stream, of the plurality of data streams, to remove the one or more idles.

9. The method of claim 8, further comprising:

multiplexing bursts from different data streams, of the plurality of data streams, to create multiplexed data streams; and

transmitting the multiplexed data streams.

10. The method of claim 8, where the buffer associated with the data stream comprises an asynchronous buffer, the method further comprising:

writing the bursts, including the at least one burst, after packing the at least one burst, to the asynchronous buffer;

reading the bursts, including the at least one burst, from the asynchronous buffer; and

transmitting the read bursts from the one or more devices.

11. The method of claim 10, where writing the bursts comprises:

writing the bursts to the asynchronous buffer at a first frequency, and

where reading the bursts comprises;

reading the bursts from the asynchronous buffer at a second frequency, the first frequency and the second frequency differing.

12. The method of claim 10, where the buffer threshold value and a current pointer, that identifies next bursts to be read, are associated with the asynchronous buffer;

and where packing the at least one burst includes;

comparing the current pointer to the buffer threshold value, and not packing the at least one burst when the current pointer is less than the buffer threshold value.

13. The method of claim 12, further comprising:

packing the at least one burst to remove the one or more idles from the at least one burst when the current pointer is not less than the buffer threshold value.

14. The method of claim 10, where writing the bursts includes:

storing data, from the at least one burst, in the asynchronous buffer without removing any of the one or more idles when the current capacity of the buffer is less than the buffer threshold value.

15. The method of claim 14, where packing the at least one burst occurs during a first clock cycle, and

where storing the data, from the at least one burst, occurs during a second clock cycle that is subsequent to the first clock cycle.

16. The method of claim 10, where packing the at least one burst and writing the data from the at least one burst occurs during two clock cycles.

17. A network device comprising:

a plurality of physical interface cards (PICs) coupled to a packet processor of the network device, each physical interface card (PIC), of the PICs, being to;

receive, via the packet processor, packets from a plurality of packet streams,

partition the packets into bursts,

multiplex the bursts,

determine that at least one burst, from a packet stream, of the plurality of packet streams, includes one or more idles,

compare a current capacity of a buffer associated with the packet stream to a buffer threshold value;

pack, when the current capacity of the buffer is not less than the buffer threshold value, the at least one burst with data from another packet stream, of the plurality of packet streams, to remove the one or more idles, and

transmit the multiplexed bursts and the packed at least one burst.

18. A transmitter connected between a first bus and a second bus, the transmitter comprising:

a plurality of schedulers, implemented at least partially in hardware, to;

receive data on the first bus, and

partition the data into bursts;

a buffer, associated with the plurality of schedulers, to;

determine that at least burst, of the bursts, includes one or more idles,

determine, based on a comparison, to pack the at least one burst with data to remove at least one idle from the at least one burst,

the comparison comparing a current capacity of the buffer to a buffer threshold value; and

merge logic, associated with the plurality of schedulers, to;

pack, based on the determining to pack the at least one burst, the at least one burst with the data to remove the at least one idle from the at least one burst, and

transmit the bursts, including the at least one burst, without the at least one idle, on the second bus,

the second bus being narrower and faster than the first bus.

19. The transmitter of claim 18, where the merge logic includes:

one or more multiplexers to selectively pack the at least one burst with data based on a capacity of a memory associated with transmitting the bursts.

20. The transmitter of claim 18, where the merge logic is further to:

write the bursts, including the packed at least one burst, in an asynchronous buffer at a first frequency; and

read the bursts, including the packed at least one burst, from the asynchronous buffer at a second frequency,

the first frequency differing from the second frequency.

21. The transmitter of claim 20, where the first frequency is greater than the second frequency.