Server-centric high performance network architecture for modular data centers

Patent metadata

Field	Value
Publication number	US-9674082-B2
Application number	US-201615040819-A
Country	US
Kind code	B2
Filing date	Feb 10, 2016
Priority date	Jan 9, 2009
Publication date	Jun 6, 2017
Grant date	Jun 6, 2017

Abstract

Disclosed are systems and methods for network architecture that is a server-centric network architectural design.

First claim

What is claimed is:

1. A method comprising:
determining multiple flow paths between a source server and a destination computing device of a network architecture;
sending, from the source server, one or more first probing packets via a first flow path of the multiple flow paths;
sending, from the source server, one or more second probing packets via a second flow path of the multiple flow paths;
receiving, at the source server, one or more first response packets, the one or more first response packets comprising first bandwidth information associated with the first flow path;
receiving, at the source server, one or more second response packets, the one or more second response packets comprising second bandwidth information associated with the second flow path;
determining, at the source server and based at least in part on at least one of the first bandwidth information or the second bandwidth information, a flow path of the multiple flow paths for packet flow to traverse; and
selecting, at the source server, the flow path for the packet flow to traverse.

2. The method of claim 1 , further comprising sending the packet flow to the destination computing device using the flow path.

3. The method of claim 1 , wherein determining the flow path comprises determining whether to buffer packets for upcoming flows of packets.

4. The method of claim 1 , wherein selecting the flow path comprises selecting the flow path from parallel flow paths of the multiple flow paths.

5. The method of claim 1 , wherein selecting the flow path comprises determining that the flow path has maximum available bandwidth among flow paths included in the multiple flow paths.

6. The method of claim 1 , wherein the determining the flow path of the multiple flow paths for the packet flow to traverse comprises determining, based at least in part on the first bandwidth information and the second bandwidth information, that the first flow path has roughly a same amount of available bandwidth as the second flow path; and
wherein selecting the flow path comprises, in response to determining that the first flow path has roughly the same amount of available bandwidth as the second flow path:
determining that the first flow path has a shorter length than the second flow path; and
based at least in part on the first flow path having the shorter length than the second flow path, selecting the first flow path for the packet flow to traverse.

7. The method of claim 1 , further comprising:
prior to determining the flow path of the multiple flow paths for the packet flow to traverse, selecting a default path for the packet flow to traverse;
sending the packet flow to the destination computing device via the default path; and
in response to selecting the flow path for the packet flow to traverse, sending the packet flow to the destination computing device using the flow path.

8. The method of claim 1 , further comprising:
periodically probing the multiple flow paths, the probing comprising:
sending, from the source server, one or more probing packets via each of the multiple flow paths; and
receiving, at the source server, one or more response packets, the one or more response packets comprising periodic bandwidth information associated with each of the multiple flow paths;

receiving, at the source server, a path failure message indicating that the flow path selected for the packet flow to traverse is unavailable for the packet flow to traverse; and
selecting, at the source server, another flow path from the multiple flow paths for the packet flow to traverse based at least in part on the periodic bandwidth information.

9. A system comprising:
one or more processors;
memory communicatively coupled to the one or more processors;
computer-readable instructions stored in the memory and that, when executed by the one or more processors, cause the one or more processors to perform acts comprising:
determining a first flow path between a source sever and a first set of servers;
determining a second flow path between the source server and a second set of servers;
sending, from the source server, one or more first probing packets via the first flow path;
sending, from the source server, one or more second probing packets via the second flow path;
receiving, at the source server, one or more first response packets, the one or more first response packets comprising first available bandwidth information associated with the first flow path;
receiving, at the source server, one or more second response packets, the one or more second response packets comprising second available bandwidth information associated with the second flow path;
determining, at the source server and based at least in part on at least one of the first available bandwidth information or the second available bandwidth information, that the first flow path has a first amount of available bandwidth that is greater than a second amount of available bandwidth of the second flow path; and
selecting, at the source server, the first flow path for a packet flow to traverse based at least partly on the determining that the first amount of available bandwidth of the first flow path is greater than the second amount of available bandwidth of the second flow path.

10. The system of claim 9 , wherein the acts further comprise encoding the first flow path into packet headers associated with the packet flow.

11. The system of claim 9 , wherein selecting the first flow path comprises determining that the first flow path and the second flow path are parallel flow paths.

12. The system of claim 9 , wherein selecting the first flow path comprises selecting the first flow path and the second flow path for the packet flow to traverse.

13. The system of claim 9 , wherein selecting the first flow path comprises determining whether to buffer packets for upcoming flows of packets.

14. The system of claim 9 , wherein selecting the first flow path comprises determining that the first flow path is shorter than the second flow path.

15. The system of claim 9 , wherein the determining that the first flow path has a first amount of available bandwidth that is greater than a second amount of available bandwidth of the second flow path comprises:
analyzing a first bandwidth field of the one or more first response packets to identify the first amount of available bandwidth of the first flow path, the first bandwidth field being populated by a destination computing device; and
analyzing a second bandwidth field of the one or more second response packets to identify the second amount of available bandwidth of the second flow path, the second bandwidth field being populated by the destination computing device.

16. One or more computer-readable media storing instructions that, when executed by one or more processors, cause the one or more processors to perform acts comprising:
determining multiple flow paths between a source server and a destination computing device of a network architecture;
sending, from the source server, one or more first probing packets via a first flow path of the multiple flow paths;
sending, from the source server, one or more second probing packets via a second flow path of the multiple flow paths;
receiving, at the source server, one or more first response packets, the one or more first response packets comprising first bandwidth information associated with the first flow path;
receiving, at the source server, one or more second response packets, the one or more second response packets comprising second bandwidth information associated with the second f

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Guo Chuanxiong
• Wu Haitao
• Lu Songwu
• Shi Yunfeng
• Lv Guohan
• Zhang Danfeng
• Li Dan
• Zhang Yongguang

Classifications

• H04L45/745

  Address table lookup; Address filtering · CPC title

• H04L49/10

  characterised by the switching fabric construction · CPC title

• H04L45/24

  Multipath · CPC title

• H04L45/125

  based on throughput or bandwidth · CPC title

• H04L47/626

  channel conditions · CPC title