REDUNDANT APPLICATION NETWORK APPLIANCES USING A LOW LATENCY LOSSLESS INTERCONNECT LINK

• US 20090063893A1
• Filed: 04/11/2008
• Published: 03/05/2009
• Est. Priority Date: 08/28/2007
• Status: Active Grant

• Current Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Original Assignee: Rohati Systems, Inc. (Cisco Systems, Inc.)
• Inventors: Patra, Abhijit, Prabhu, Kirti, Thakar, Anant, Bagepalli, Nagaraj, Gandhi, Prashant
• Granted Patent: US 7,895,463 B2
• Time in Patent Office: Days
• US Class Current: 714/4
• CPC Class Codes: H04L 47/20 Traffic policing
• CPC Class Codes: H04L 63/02 for separating internal from external traffic, e.g. firewalls
• CPC Class Codes: H04L 63/0428 wherein the data content is protected, e.g. by encrypting or encapsulating the payload
• CPC Class Codes: H04L 63/166 at the transport layer
• CPC Class Codes: H04L 63/205 involving negotiation or determination of the one or more network security mechanisms to be used, e.g. by negotiation between the client and the server or between peers or by selection according to the capabilities of the entities involved negotiation of communication capabilities H04L69/24
• CPC Class Codes: H04L 69/16 Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
• CPC Class Codes: H04L 69/161 Implementation details of TCP/IP or UDP/IP stack architecture; Specification of modified or new header fields
• CPC Class Codes: H04L 69/321 Interlayer communication protocols or service data unit [SDU] definitions; Interfaces between layers
• CPC Class Codes: H04L 9/3242 involving keyed hash functions, e.g. message authentication codes [MACs], CBC-MAC or HMAC
• CPC Class Codes: Y10T 70/5827 Axially movable clutch

Abstract

Redundant application network appliances using a low latency lossless interconnect link are described herein. According to one embodiment, in response to receiving at a first network element a packet of a network transaction from a client over a first network for accessing a server of a datacenter, a layer 2 network process is performed on the packet and a data stream is generated. The data stream is then replicated to a second network element via a layer 2 interconnect link to enable the second network element to perform higher layer processes on the data stream to obtain connection states of the network transaction. In response to a failure of the first network element, the second network element is configured to take over processes of the network transaction from the first network element using the obtained connection states without user interaction of the client. Other methods and apparatuses are also described.

Claims

1. A method performed by a network element, the method comprising:

in response to receiving at a first network element a packet of a network transaction from a client over a first network for accessing a server of a datacenter having a plurality of servers over a second network, performing a layer 2 network process on the packet and generating a data stream within the first network element;

performing a layer-3 handshaking with a second network element via a lossless interconnect link coupling the first and second network elements; and

replicating the data stream to the second network element via the lossless interconnect link to synchronize layer-4 states between the first and second network elements and to enable the second network element to perform higher layer processes on the data stream with respect to the network transaction to obtain connection states of the network transaction, wherein the second network element is configured to operate as a redundant peer to the first network element, and wherein in response to a failure of the first network element, the second network element is configured to take over processes of the network transaction from the first network element using the obtained connection states without user interaction of the client.

2. The method of claim 1, wherein the lossless interconnect link is implemented using one of an RDMA (remote direct memory access) capable and an InfiniBand compatible technology and/or lossless interconnect.

3. The method of claim 1, further comprising:

performing layers 5-7 process on the packet within the first network element to obtain layers 5-7 connection states of the network transaction;

storing the layers 5-7 connection states in a first memory of the first network element; and

synchronizing the layers 5-7 connection states with the second network element by synchronizing at least a portion of content of the first memory of the first network element and at least a portion of content of a second memory of a second network element via an predetermined low latency lossless interconnect link.

4. The method of claim 3, wherein the memories of the first network element and the second network element are synchronized via an RDMA capable and/or lossless fabric.

5. The method of claim 4, further comprising exposing at least a portion of the first memory of the first network element to the second network element such that the second network element is able to read content of the exposed at least a portion of first memory of the first network to obtain the layers 5-7 connection states.

6. The method of claim 5, wherein the first memory of the first network element is read by the second network element via a RDMA channel without involving significant processing power from the first network element.

7. The method of claim 5, wherein the second network element directly accesses the exposed at least a portion of the first memory of the first network element via an RDMA channel bypassing a TCP (transport control protocol) network stack of the first network element.

8. The method of claim 1, wherein the second network element is configured monitor health of the first network element via a dedicated monitoring channel over the lossless interconnect link to determine whether the first network element is operating properly, and wherein if it is determined that the first network element fails to operate properly, the second network element is configured to transparently takes over the network processes of the network transaction using the obtained connection states without acknowledge from the client.

9. A machine-readable medium having instructions stored therein, which when executed by a machine, cause the machine to perform a method, the method comprising:

in response to receiving at a first network element a packet of a network transaction from a client over a first network for accessing a server of a datacenter having a plurality of servers over a second network, performing a layer 2 network process on the packet and generating a data stream within the first network element;

performing a layer-3 handshaking with a second network element via a lossless interconnect link coupling the first and second network elements; and

replicating the data stream to the second network element via the lossless interconnect link to synchronize layer-4 states between the first and second network elements and to enable the second network element to perform higher layer processes on the data stream with respect to the network transaction to obtain connection states of the network transaction, wherein the second network element is configured to operate as a redundant peer to the first network element, and wherein in response to a failure of the first network element, the second network element is configured to take over processes of the network transaction from the first network element using the obtained connection states without user interaction of the client.

10. The machine-readable medium of claim 9, wherein the lossless interconnect link is implemented using one of an RDMA (remote direct memory access) capable and an InfiniBand compatible technology and/or lossless interconnect.

11. The machine-readable medium of claim 9, wherein the method further comprises:

performing layers 5-7 process on the packet within the first network element to obtain layers 5-7 connection states of the network transaction;

storing the layers 5-7 connection states in a first memory of the first network element; and

synchronizing the layers 5-7 connection states with the second network element by synchronizing at least a portion of content of the first memory of the first network element and at least a portion of content of a second memory of a second network element via an predetermined low latency lossless interconnect link.

12. The machine-readable medium of claim 11, wherein the memories of the first network element and the second network element are synchronized via an RDMA capable and/or lossless fabric.

13. The machine-readable medium of claim 12, wherein the method further comprises exposing at least a portion of the first memory of the first network element to the second network element such that the second network element is able to read content of the exposed at least a portion of first memory of the first network to obtain the layers 5-7 connection states.

14. The machine-readable medium of claim 13, wherein the first memory of the first network element is read by the second network element via a RDMA channel without involving significant processing power from the first network element.

15. The machine-readable medium of claim 13, wherein the second network element directly accesses the exposed at least a portion of the first memory of the first network element via an RDMA channel bypassing a TCP (transport control protocol) network stack of the first network element.

16. The machine-readable medium of claim 9, wherein the second network element is configured monitor health of the first network element via a dedicated monitoring channel over the lossless interconnect link to determine whether the first network element is operating properly, and wherein if it is determined that the first network element fails to operate properly, the second network element is configured to transparently takes over the network processes of the network transaction using the obtained connection states without acknowledge from the client.

17. A network element, comprising:

a network service module; and

an application service module coupled to the network service module over a switch fabric,

wherein in response to receiving a packet of a network transaction from a client over a first network for accessing a server of a datacenter having a plurality of servers over a second network, the network service module is configured to perform a layer 2 process on the packet which generates a data stream,

wherein a layer 3 handshaking process is performed over a lossless interconnect coupling the network element with a second network element,

wherein the data stream is replicated via the lossless interconnect link to the second network element as a redundant peer to synchronize layer-4 states with the redundant peer and to enable the redundant peer to perform higher layer processes on the data stream to obtain connection states of the network transaction, and

wherein in response to a failure of the network element, the redundant peer is configured to take over network processes of the network transaction using the obtained connection states without user interaction from the client.

18. The network element of claim 17, wherein the lossless interconnect link is implemented using one of an RDMA (remote direct memory access) capable and an InfiniBand compatible technology and/or lossless interconnect.

19. The network element of claim 17, wherein the network service module is configured to perform a layers 5-7 process on the packet to obtain layers 5-7 connection states of the network transaction and to store the layers 5-7 connection states in a first memory of the network element, wherein the layers 5-7 connection states are synchronized with the redundant peer by synchronizing at least a portion of content of the first memory and at least a portion of content of a second memory of redundant peer via an predetermined low latency lossless interconnect link.

20. The network element of claim 19, wherein the memories of the network element and the redundant peer are synchronized via an RDMA capable and/or lossless fabric.

21. The network element of claim 20, wherein at least a portion of the first memory of the network element is exposed to the redundant peer such that the redundant peer is able to read content of the exposed at least a portion of first memory of the network to obtain the layers 5-7 connection states.

22. The network element of claim 21, wherein the first memory of the network element is read by the redundant peer via a RDMA channel without involving significant processing power from the network element.

23. The network element of claim 21, wherein the redundant peer directly accesses the exposed at least a portion of the first memory of the network element via an RDMA channel bypassing a TCP (transport control protocol) network stack of the network element.

24. The network element of claim 17, wherein the redundant peer is configured monitor health of the network element via a dedicated monitoring channel over the lossless interconnect link to determine whether the network element is operating properly, and wherein if it is determined that the network element fails to operate properly, the redundant peer is configured to transparently takes over the network processes of the network transaction using the obtained connection states without acknowledge from the client.