Summary of the invention
The object of the invention is to overcome processor process in prior art to roll off the production line the problems such as frame message required time is long, delay jitter is large, provide a kind of to the MRP looped network redundancy treatment system that frame message processing speed is faster, the response time is shorter that rolls off the production line.
In order to realize foregoing invention object, the invention provides following technical scheme:
A kind of hardware based MRP looped network redundancy treatment system, comprises,
Exchange chip, receive the protocol massages (generally comprising the frame message LinkDown that rolls off the production line, test frame message Test, topology change frame message TopologyChange) that MRP looped network medium redundant management node MRM uploads, this protocol massages is sent to processor or hardware co-processor, the result that receiving processor or hardware co-processor return, and according to this result, medium redundant management node MRM is operated; General, this operation refers to: MRM port is transformed into FORWARDING state by BLOCKED by exchange chip 1; Control MRP looped network is opened a way and network service that is newly-built, that recover topological structure; It should be noted, the operation that above-mentioned exchange chip carries out for medium redundant management node MRM from there is no in prior art different.
Hardware co-processor, receives the frame message that rolls off the production line that described exchange chip is uploaded, and carries out response process and result (or being referred to as control command) is back to described exchange chip to this message; In the present invention, described hardware co-processor is that FPGA or NP realizes, only for receiving, processing the frame message that rolls off the production line.
Processor, the protocol massages outside the frame message that rolls off the production line that reception exchange chip is uploaded also responds process, and result is back to described exchange chip.
It should be noted, in the present invention, processor and hardware co-processor all receive data that exchange chip uploads (data refer to comprise the various messages of frame message of rolling off the production line) by data/address bus (as MII, PCIE etc.), by control bus (as SPI etc.) to exchange chip issue control signal (as, to the result of the frame message that rolls off the production line, result also can be referred to as processing command).
In some embodiments, hardware co-processor, between processor and exchange chip, both improve transmission rate, in turn saved the use of exchange chip pin.
Further, described exchange chip receive that medium redundant management node MRM uploads roll off the production line frame message time, judge that whether this frame message that rolls off the production line processed, as it is processed, abandoned, as untreated, be uploaded to described hardware co-processor.
Further, described hardware co-processor is also sent to described processor by after the processed frame packet labeling that rolls off the production line.
Further, during the excessive or hardware co-processor fault of the frame message data amount that rolls off the production line pending in described hardware co-processor, the frame message that rolls off the production line received is uploaded to described processor process.
Further, after described processor receives the frame message that rolls off the production line that hardware co-processor sends, whether the frame message that rolls off the production line described in inspection is processed, and as processed, then whether the MRP ring-network topology state after check processing is correct; As untreated, then to its response process, and result is sent to exchange chip.
The present invention provides a kind of to the MRP looped network redundancy processing method that frame message processing speed is faster, the response time is shorter that rolls off the production line simultaneously, comprises the steps:
(1) exchange chip receives the message uploaded from MRP looped network medium redundant management node MRM, judges that whether this message rolls off the production line frame message, in this way, is sent to hardware co-processor; Otherwise be uploaded to processor;
(2) hardware co-processor connects the roll off the production line frame message of response from exchange chip, and result is back to described exchange chip;
(3) exchange chip operates to recover the connection of MRP looped network to medium redundant management node MRM according to result.
Further, after described exchange chip receives the frame message that rolls off the production line, also comprise described exchange chip judge described in roll off the production line the step whether frame message processed; As it is processed, abandoned, as untreated, be uploaded to described hardware co-processor.
Further, roll off the production line described in described hardware co-processor response process after frame message, also comprising described hardware co-processor will be sent to the step of described processor after the processed frame packet labeling that rolls off the production line.
Further, roll off the production line after frame message described in described hardware co-processor receives, when also comprising the excessive or hardware co-processor fault of the frame message data amount that rolls off the production line pending in hardware co-processor, the frame message that rolls off the production line received is uploaded to the step of described processor process.
Further, also comprise, after described processor receives the frame message that rolls off the production line that hardware co-processor sends, whether the frame message that rolls off the production line described in inspection is processed, and as processed, then whether the MRP ring-network topology state after check processing is correct; As untreated, then to its response process, and result is sent to the step of exchange chip.
compared with prior art, beneficial effect of the present invention: the frame message that rolls off the production line that hardware based looped network redundancy treatment system provided by the invention is uploaded by arranging hardware co-processor response processing and exchanging chip, because the response time of hardware co-processor is short, processing speed is fast, delay jitter is little, greatly can reduce the time of recovering network service like this, reduce packet loss, effectively can improve switch performance.
Embodiment
Below in conjunction with drawings and the specific embodiments, the present invention is described in further detail.But this should be interpreted as that the scope of the above-mentioned theme of the present invention is only limitted to following embodiment, all technology realized based on content of the present invention all belong to scope of the present invention.
embodiment 1:as shown in Figure 2, the present embodiment provides a kind of hardware based MRP looped network redundancy treatment system, comprises,
Exchange chip 1, receive the protocol massages that MRP looped network medium redundant management node MRM uploads, this protocol massages is sent to processor 2 or hardware co-processor 3, the result that receiving processor 2 or hardware co-processor 3 return, and according to this result, medium redundant management node MRM is operated; General, this operation refers to: MRM port is transformed into FORWARDING state by BLOCKED by exchange chip 1; Control MRP looped network is opened a way and network service that is newly-built, that recover topological structure.
Hardware co-processor 3, receives the frame message LinkDown that rolls off the production line that described exchange chip 1 is uploaded, and carries out response process and result is back to described exchange chip 1 to this message; In the present embodiment, described hardware co-processor 3 realizes for FPGA, only for receiving, processing the frame message LinkDown that rolls off the production line, and does not process other messages.
Processor 2, the protocol massages (as test frame message Test, topology change frame message TopologyChange) outside the frame message LinkDown that rolls off the production line that reception exchange chip 1 is uploaded also responds process, result is back to described exchange chip 1.
It should be noted, in the present invention, processor 2 and hardware co-processor 3 are all passed through data/address bus (as MII, PCIE etc.) and are received data that exchange chip 1 uploads (data refer to comprise the various protocol massages of frame message LinkDown of rolling off the production line), control signal (e.g., to the result of the frame message LinkDown that rolls off the production line) is issued to exchange chip 1 by control bus (as SPI etc.).
Further, in the present embodiment, described exchange chip 1 receive that medium redundant management node MRM uploads roll off the production line frame message LinkDown time, first judge that whether this frame message LinkDown that rolls off the production line is processed, as it is processed, abandoned, as untreated, be uploaded to described hardware co-processor.
And in other embodiment, described exchange chip 1 is after receiving the frame message LinkDown that rolls off the production line, do not judge that whether it is processed, but be directly uploaded to hardware co-processor 3, hardware co-processor 3 processes its response according to normal procedure, the effect that after can playing the process to MRP looped network like this, state is rechecked.
Further, described hardware co-processor 3 is sent to described processor 2 after also being marked by the processed frame message LinkDown that rolls off the production line, and mark refers to and is labeled as processed by the described frame message LinkDown that rolls off the production line.
Further, during excessive or hardware co-processor 3 fault of the frame message LinkDown data volume that rolls off the production line pending in described hardware co-processor 3, the frame message LinkDown that rolls off the production line received is uploaded to described processor 2 and processes.
After described processor 2 receives the frame message LinkDown that rolls off the production line that hardware co-processor 3 sends, whether the frame message LinkDown that rolls off the production line described in inspection is processed, and as processed, then whether the MRP ring-network topology state after check processing is correct; As untreated, then to its response process, and result is sent to exchange chip 1.
embodiment 2:as shown in Figure 6, in the present embodiment, hardware co-processor 3 is between processor 2 and exchange chip 1, both improve transmission rate, in turn save the use of exchange chip pin, under this pattern, the message received (referring to roll off the production line frame message LinkDown, test frame message Test, topology change frame message TopologyChange) all need be reached processor 2 by hardware co-processor 3 by exchange chip 1.
embodiment 3:as shown in Figure 3, the present embodiment provides a kind of to the MRP looped network redundancy processing method that frame message LinkDown processing speed is faster, the response time is shorter that rolls off the production line, and comprises the steps:
S100: exchange chip 1 receives the message uploaded from MRP looped network medium redundant management node MRM, judges that whether this message rolls off the production line frame message LinkDown, in this way, is sent to hardware co-processor 3; Otherwise be uploaded to processor 2, in general, the message that medium redundant management node MRM uploads comprises the frame message LinkDown that rolls off the production line, test frame message Test, topology change frame message TopologyChange; Namely in the present embodiment, the frame message LinkDown that rolls off the production line can be uploaded to hardware co-processor 3 by exchange chip 1, and test frame message Test, topology are changed frame message TopologyChange and be uploaded to processor 2.
S200: hardware co-processor 3 meets the roll off the production line frame message LinkDown of response from exchange chip 1, result is back to described exchange chip 1;
S300: exchange chip 1 operates to recover MRP looped network according to result to medium redundant management node MRM and connects.
embodiment 4:as shown in Figure 4, the present embodiment provides a kind of to the MRP looped network redundancy processing method that frame message LinkDown processing speed is faster, the response time is shorter that rolls off the production line, and comprises the steps:
S110: exchange chip 1 receives the message uploaded from MRP looped network medium redundant management node MRM, judges whether this message rolls off the production line frame message LinkDown; As no, be then uploaded to processor 2;
In this way, then step S111 is entered;
S120: whether the frame message LinkDown that rolls off the production line described in judgement processed; As it is processed, abandoned, as untreated, be uploaded to described hardware co-processor 3.
S210: hardware co-processor 3 meets the roll off the production line frame message LinkDown of response from exchange chip 1, result is back to described exchange chip 1;
Meanwhile, the described hardware co-processor 3 frame message LinkDown that will roll off the production line reaches described processor 2; Comprise two kinds of situations, wherein a kind of situation is: hardware co-processor 3 is sent to described processor 2 after being marked by the processed frame message LinkDown that rolls off the production line; Another situation is: when the frame message LinkDown data volume that rolls off the production line pending in hardware co-processor 3 is excessive or hardware co-processor 3 fault time, the undressed frame message LinkDown that rolls off the production line received is uploaded to described processor 2 and processes.
S220: after described processor 3 receives the frame message LinkDown that rolls off the production line that hardware co-processor 3 sends, whether the frame message LinkDown that rolls off the production line described in inspection is processed, as processed, then the MRP ring-network topology state after check processing whether correct (changing frame message TopologyChange by topology to detect); As untreated, then to its response process, and result is sent to exchange chip 1.
S300: exchange chip 1 operates to recover MRP looped network according to the result received to medium redundant management node MRM and connects.
embodiment 5:as shown in Figure 5, the present embodiment provides a kind of to the MRP looped network redundancy processing method that frame message LinkDown processing speed is faster, the response time is shorter that rolls off the production line, and comprises the steps:
S100: exchange chip 1 receives the message uploaded from MRP looped network medium redundant management node MRM, judges that whether this message rolls off the production line frame message LinkDown, in this way, is sent to hardware co-processor 3; Otherwise be uploaded to processor 2;
S210: hardware co-processor 3 meets the roll off the production line frame message LinkDown of response from exchange chip 1, result is back to described exchange chip 1;
Meanwhile, the described hardware co-processor 3 frame message LinkDown that will roll off the production line reaches described processor 2; Comprise two kinds of situations, wherein a kind of situation is: hardware co-processor 3 is sent to described processor 2 after being marked by the processed frame message LinkDown that rolls off the production line; Another situation is: when the frame message LinkDown data volume that rolls off the production line pending in hardware co-processor 3 is excessive or hardware co-processor 3 fault time, the undressed frame message LinkDown that rolls off the production line received is uploaded to described processor 2 and processes.
S220: after described processor 3 receives the frame message LinkDown that rolls off the production line that hardware co-processor 3 sends, whether the frame message LinkDown that rolls off the production line described in inspection is processed; As untreated, then to its response process, and result is sent to exchange chip 1.As processed, then whether the MRP ring-network topology state after check processing is correct, as MRP ring-network topology state is correct, then terminates; As incorrect in MRP ring-network topology state, then response process is carried out to message;
S300: exchange chip 1 operates to recover MRP looped network according to the result received to medium redundant management node MRM and connects.