Data center network using circuit switching

Patent metadata

Field	Value
Publication number	US-9270484-B2
Application number	US-201213355575-A
Country	US
Kind code	B2
Filing date	Jan 23, 2012
Priority date	Jan 23, 2012
Publication date	Feb 23, 2016
Grant date	Feb 23, 2016

Abstract

A circuit-based digital communications network is provided for a large data center environment that utilizes circuit switching in lieu of packet switching in order to lower the cost of the network and to gain performance efficiencies. A method for transmitting data in such a network comprises sending a setup request for a path for transmitting the data to a destination node and then speculatively sending the data to the destination node before the setup request is completed.

First claim

What is claimed:

1. A data center comprising:
a first switch of a first network, wherein the first switch is configured to use circuit-switching for routing data between an inward-facing portion of the first network and an outward-facing portion of the first network, wherein the data is formatted in a repeating frame format, and wherein each frame of the repeating frame format comprises a plurality of control slots and a plurality of data slots; and
a plurality of servers communicatively coupled to the first switch via the inward-facing portion of the first network, wherein the plurality of servers are interconnected to each other on a second network, and wherein the second network is configured to use bandwidth reservation for transporting data to and from the first switch of the first network,
wherein the data center is capable of sending a setup request for a path for transmitting the data to a destination node in the first network, and speculatively sending, separate from the setup request, the data through the first network to the destination node before the setup request is completed, and wherein the destination node does not respond to the setup request if it accepts the setup request and receives the speculatively sent data, and wherein the destination node responds to the setup request, the setup request is killed, and a tear down message is received if the destination node does not accept the setup request.

2. The data center of claim 1 , wherein the first switch comprises a crossbar.

3. The data center of claim 1 , wherein the first switch is communicatively coupled to the plurality of servers via a network interface controller (NIC).

4. The data center of claim 3 , wherein the first switch comprises at least one field programmable gate array (FPGA) and wherein the NIC comprises at least one application-specific integrated circuit (ASIC).

5. The data center of claim 1 , further comprising a network operations center (NOC) coupled to the outward-facing portion of the first network.

6. The data center of claim 1 , further comprising a second switch communicatively coupled to at least the first switch via redundant interconnections.

7. The data center of claim 6 , wherein the redundant interconnections are configured to provide load balancing.

8. The data center of claim 1 , wherein:
each server in the plurality of servers is assigned a unique destination address based on a nodal division number, a nodal group number, and a node number;
wherein the nodal division number corresponds to a port number on the plurality of servers for the outward-facing portion of the first network; and
further wherein the nodal group number corresponds to a port number on at least one server for the inward-facing portion of the first network.

9. The data center of claim 1 , wherein the first switch comprises a plurality of line cards orthogonally interconnected to a plurality of crossbar cards.

10. The data center of claim 9 , wherein each crossbar card from among the plurality of crossbar cards comprises at least one FPGA.

11. The data center of claim 10 , wherein the plurality of line cards and the plurality of crossbar cards are interconnected via a midplane.

12. The data center of claim 1 , wherein each of the plurality of control slots comprises a first field and a second field, wherein the first field comprises a control message and the second field comprises a header for the plurality of data slots.

13. The data center of claim 1 , wherein each frame of the repeating frame format further comprises a plurality of tag bits.

14. The data center of claim 13 , wherein the plurality of tag bits are configured for clock recovery by selecting combinations of tag bits that guarantee at least one data transition over a preselected number of data bits.

15. A system comprising: a plurality of servers, wherein each server comprises a network interface controller (NIC) for interconnecting the plurality of servers using a communication format wherein data is transported in cells; and at least one switch of a first network, wherein the at least one switch is communicatively coupled to at least one of the plurality of servers via the NIC, wherein the plurality of servers are interconnected to each other on a second network, wherein the at least one switch is configured to use circuit-switching for routing data to the at least one of the plurality of servers, wherein the data routed to the at least one of the plurality of servers is in a repeating frame format, wherein each frame of the repeating frame format comprises a plurality of control slots and a plurality of data slots, and wherein the second network is configured to use bandwidth reservation for transporting data to and from the at least one switch of the first network, wherein the system is capable of sending a setup request for a path for transmitting the data to a destination node in the first network, and speculatively sending, separate from the setup request, the data through the first network to the destination node before the setup request is completed, and wherein the destination node does not respond to the setup request if it accepts the setup request and receives the speculatively sent data, and wherein the destination node responds to the setup request, the setup request is killed, and a tear down message is received if the destination node does not accept the setup request.

16. The system of claim 15 , wherein the at least one switch comprises at least one crossbar card.

17. The system of claim 16 , further comprising a plurality of crossbar cards.

18. A method for transmitting data, the method comprising: sending a setup request for a path for transmitting data to a destination node in a circuit-switched digital communications network; and speculatively sending, separate from the setup request, the data through the circuit-switched digital communications network to the destination node before the setup request is completed, wherein the data is sent in a repeating frame format, wherein each frame of the repeating frame format comprises a plurality of data slots and a plurality of control slots, and wherein bandwidth reservation is used for transporting the data in the circuit-switched digital communications network, and wherein the destination node does not respond to the setup request if it accepts the setup request and receives the speculatively sent data, and wherein the destination node responds to the setup request, the setup request is killed, and a tear down message is received if the destination node does not accept the setup request.

19. The method of claim 18 , wherein a tear down message is received if network capacity of the circuit-switched digital communications network is exceeded.

20. The method of claim 18 , wherein sending the data through the circuit-switched digital communications network is completed unless a tear down message is received.

Assignees

• Thacker Charles P
• Nowatzyk Andreas G
• Yu Fang
• Rodeheffer Thomas L
• Microsoft Technology Licensing Llc

Inventors

• Thacker Charles P
• Nowatzyk Andreas G
• Yu Fang
• Rodeheffer Thomas L

Classifications

• H04L41/0893

  Assignment of logical groups to network elements · CPC title

• H04L41/0896

  Bandwidth or capacity management, i.e. automatically increasing or decreasing capacities (flow or congestion control using dynamic resource allocation, e.g. in-call renegotiation, H04L47/76) · CPC title

• H04L12/437 Primary

  Ring fault isolation or reconfiguration {(for SDH/SONET ring networks H04J3/085)} · CPC title

• H04L41/0823

  characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability (for optimising operational conditions of wireless networks H04W24/02) · CPC title

• H04L49/356

  for storage area networks · CPC title