IPC분류정보
국가/구분 |
United States(US) Patent
등록
|
국제특허분류(IPC7판) |
|
출원번호 |
US-0891992
(2004-07-15)
|
등록번호 |
US-8228931
(2012-07-24)
|
발명자
/ 주소 |
- Wallace, Raymond Bruce
- Chmara, Thomas P.
- Stark, Ryan
|
출원인 / 주소 |
|
대리인 / 주소 |
Christopher & Weisberg, P.A.
|
인용정보 |
피인용 횟수 :
5 인용 특허 :
20 |
초록
▼
An architecture that enables distributed storage capacity to be shared as though accesses were being made to a Fiber Channel Switch is described. The distributed capacity may include Fiber Channel devices and thus a method for addressing Fiber Channel devices coupled in a resilient ring has been pro
An architecture that enables distributed storage capacity to be shared as though accesses were being made to a Fiber Channel Switch is described. The distributed capacity may include Fiber Channel devices and thus a method for addressing Fiber Channel devices coupled in a resilient ring has been provided. Expected Fiber Channel behavior and reliability are attained in the architecture of this invention through the application of bandwidth provisioning and extended reach mechanisms at each node in the ring. The bandwidth provisioning mechanisms ensure that access is provided to the communication links of the ring for each node according to a defined bandwidth distribution. The extended reach mechanisms of the invention ensure that the standard Fiber Channel packet delivery assumptions are satisfied. The solution exploits native ring characteristics to deliver customer separation and management of the resulting storage network. Forwarding Error Correction techniques can be used to ensure that a Bit Error Rate of the Fiber Channel communications on the ring is comparable to a Bit Error Rate of a typical Fiber Channel network.
대표청구항
▼
1. A distributed virtual storage switch comprising: a plurality of metro spaced nodes, wherein the metro spaced nodes are coupled in a ring;a first one of the nodes being in communication with a first storage device via a first storage switch; anda second one of the nodes being in communication with
1. A distributed virtual storage switch comprising: a plurality of metro spaced nodes, wherein the metro spaced nodes are coupled in a ring;a first one of the nodes being in communication with a first storage device via a first storage switch; anda second one of the nodes being in communication with a second storage device via a second storage switch;a third one of the nodes accessible by the first and second nodes using logical point to point connections between nodes on the ring, the third node allocating bandwidth to the first and second nodes by placing a number of empty frames in transmissions from the third node, the empty frames being accessible by, and replaced by data by, the first and second nodes, the number of empty frames being in a fixed proportion to a number of non-empty data frames in transmissions from the third node; andthe plurality of nodes of the ring providing a virtual storage switch for reaching the first and second storage devices. 2. The distributed virtual switch of claim 1, wherein a physical link capacity between at least two of the nodes on the ring is allocated to one of the plurality of logical point to point connections according to a bandwidth provisioning mechanism. 3. The distributed virtual switch of claim 2 wherein the bandwidth provisioning mechanism is programmable. 4. The distributed virtual switch of claim 2, wherein the bandwidth provisioning mechanism selects one of the plurality of logical point to point connections to use the physical link capacity according to a relative service level associated with the logical point to point connection. 5. The distributed virtual switch of claim 1, wherein one or more users may be associated with the plurality of nodes, and wherein the distributed virtual switch further comprises an addressing mechanism enabling separation of communications between the one or more users. 6. The distributed virtual switch of claim 5, wherein the one or more of the users are associated with one or more customers. 7. The distributed virtual switch of claim 6 wherein the addressing mechanism is programmable. 8. The distributed virtual switch of claim 1 further comprising an extended reach mechanism enabling point to point logical storage protocol exchanges over the ring. 9. The distributed virtual switch of claim 8 wherein the extended reach mechanism is programmable. 10. The distributed virtual storage switch of claim 1, wherein the first one of the nodes asserts a flow control signal when the first one of the nodes lacks a sufficient number of buffers available to receive a packet from another node. 11. A storage node comprising: a local interface for exchanging data with one or more coupled storage devices; anda ring interface for exchanging communications with at least one metro connected node, the communications having a message forwarding protocol corresponding to a storage protocol of the storage devices, the message forwarding protocol prepending a forwarding header to storage protocol packets to facilitate the message forwarding, the storage protocol being a Fiber Channel protocol, the forwarding header including a Transparent Domain Interconnect (TDI) address, and the storage node allocating bandwidth to the at least one metro connected node by placing empty frames in transmissions from the storage node, the empty frames being accessible by, and replaced by data by, the at least one metro connected node, the ring interface including a bandwidth provisioning mechanism for controlling access to a link coupling the storage node to the at least one metro connected node, the bandwidth provisioning mechanism including a device for storing a value associated with a bandwidth available to the storage node, the value for controlling a ratio of packets to be locally sourced by the storage node versus packets to be transmitted through the storage node on the ring interface. 12. The storage node of claim 11 wherein the value is programmable. 13. The storage node of claim 11, wherein the ring interface further comprises an extended reach mechanism for simulating a packet delivery behavior of the storage protocol using the message forwarding protocol. 14. The storage node of claim 13 wherein the extended reach mechanism includes flow control logic to selectively block transmissions by the storage nodes onto the at least one metro connected node depending upon available buffer capacity at the at least one node. 15. The storage node of claim 14 wherein the flow control logic is programmable. 16. The storage node of claim 11, wherein the addressing mechanism is used to provide customer separation. 17. The storage node of claim 14, wherein the storage protocol is a Fiber Channel protocol and the forwarding header is selected in accordance with the Fiber Channel Arbitrated Loop protocol. 18. The storage node of claim 11, wherein the storage protocol is a Fiber Channel protocol and the forwarding header is includes a switched Fiber Channel. 19. The storage node of claim 11, further comprising an error correction mechanism. 20. The storage node of claim 11, wherein a number of the empty frames is in a fixed proportion to a number of non-empty data frames in transmissions from the storage node. 21. A distributed virtual switch comprising: a plurality of nodes interconnected in a ring, at least one of the nodes being a storage node operating according to a storage protocol; anda packet exchanger at each of the nodes, each packet exchanger exchanging packets with the storage node, the packets comprising: storage protocol communications in accordance with the storage protocol: anda ring forwarding header pre-pending the storage protocol communications, the ring forwarding header being a Transparent Domain Identifier (TDI); andeach packet exchanger further translating a first packet delivery behavior of the ring into a second packet delivery behavior of the storage protocol, a first one of the nodes allocating bandwidth to at least one other node by placing a number of empty frames in transmissions from the first node, the empty frames being accessible by, and replaced by data by, the at least one other node, the number of empty frames being in a fixed proportion to a number of non-empty data frames in transmissions from the first node. 22. The distributed virtual switch of claim 21 wherein the storage protocol is a Fiber Channel protocol. 23. The distributed switch of claim 21 wherein the ring forwarding header is also used to encapsulate Internet Protocol (IP) communications, thereby enabling both IP protocol and storage protocol devices to utilize the ring. 24. The distributed switch of claim 21 wherein the ring forwarding header comprises a Fiber Channel Switch address. 25. The distributed switch of claim 21 wherein the ring forwarding header is selected in accordance with a Fiber Channel Arbitrated Loop protocol. 26. A method of transferring storage communications associated with a storage protocol between remotely distributed nodes coupled in a ring, the method including: translating a packet delivery behavior of the ring into an expected packet delivery behavior of the storage protocol so that it appears that the nodes on the ring are point to point coupled, the translating including encapsulating the storage communications in a message forwarding header, the message forwarding header being a Transparent Domain Identifier (TDI); andallocating bandwidth, by a first one of the remotely distributed nodes, to at least one other remotely distributed node by placing a number of empty frames in transmission from the first one of the remotely distributed nodes, the empty frames being accessible by, and replaced by data by, the at least one other remotely distributed node, the number of empty frames being in a fixed proportion to a number of non-empty data frames in transmissions from the first one of the remotely distributed nodes. 27. The method of claim 26 further including the step of controlling a bandwidth allocated to each of the nodes in the ring. 28. The method of claim 26 wherein at least one of the nodes in the ring is a node operating according to an Internet Protocol (IP), and wherein the method further includes the step of encapsulating internet protocol communications using the TDI. 29. The method of claim 26 wherein the message forwarding header is selected according to a Fiber Channel Arbitrated Loop (FC-AL) protocol. 30. The method of claim 26 wherein the message forwarding header includes a Fiber Channel switch address. 31. The method according to claim 26 further including the steps of identifying a bit error rate in communications between two nodes in the ring and selectively altering transmission paths between the two nodes in response to the identified bit error rate.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.