Performance and power optimized computer system architectures and methods leveraging power optimized tree fabric interconnect

• US 9,876,735 B2
• Filed: 09/15/2011
• Issued: 01/23/2018
• Est. Priority Date: 10/30/2009
• Status: Active Grant

• Current Assignee: Calxeda Inc.
• Original Assignee: III Holdings 2, LLC (Intellectual Ventures LLC)
• Inventors: Davis, Mark, Borland, David
• Primary Examiner(s): Jiang, Charles C
• Assistant Examiner(s): Soe, Kyaw Z
• Application Number:
  US13/234,054
• Publication Number:
  US 20120096211A1
• Time in Patent Office: 2,322 Days
• Field of Search: 709100, 709305, 709314, 709103, 713300, 713324
• CPC Class Codes: G06F 13/4282 on a serial bus, e.g. I2C bus, SPI bus on daisy chain buses G06F13/4247
• CPC Class Codes: H04L 47/125 by balancing the load, e.g. traffic engineering
• CPC Class Codes: H04L 49/10 characterised by the switching fabric construction
• CPC Class Codes: H04L 49/101 using crossbar or matrix
• CPC Class Codes: H04L 49/111 Switch interfaces, e.g. port details
• CPC Class Codes: H04L 49/351 for local area network [LAN], e.g. Ethernet switches
• CPC Class Codes: H04L 49/40 Constructional details, e.g. power supply, mechanical construction or backplane
• CPC Class Codes: Y02B 70/10 Technologies improving the efficiency by using switched-mode power supplies [SMPS], i.e. efficient power electronics conversion e.g. power factor correction or reduction of losses in power supplies or efficient standby modes
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

A performance and power optimized computer system architecture and method leveraging power optimized tree fabric interconnect are disclosed. One embodiment builds low power server clusters leveraging the fabric with tiled building blocks while another embodiment implements storage solutions or cooling solutions. Yet another embodiment uses the fabric to switch non-Ethernet packets, switch multiple protocols for network processors and other devices.

Claims

1. A computing device comprising:

a plurality of server nodes interconnected by a plurality of fabric links to form a server fabric, wherein the plurality of server nodes are grouped on a server board; and

one or more Ethernet escapes from the server fabric, wherein the one or more Ethernet escapes are configured to connect the server fabric to one or more ports external to the server fabric, and wherein the one or more Ethernet escapes are further configured to be pulled at any node in the plurality of server nodes to conserve bandwidth and power;

wherein a speed of each of the plurality of fabric links is configured to be dynamically adjustable based on a utilization of a particular fabric link to optimize power; and

wherein each server node that is grouped on the server board includes;

a processor;

a memory;

an input/output circuit; and

an internal fabric switch, wherein the internal fabric switch is located on the server node and is configured to connect the server node to the server fabric and to switch information in the server fabric.

2. The computing device of claim 1, further comprising an aggregate of a set of server boards, wherein each of the server boards includes one or more of the plurality of server nodes, and wherein the set of server boards are interconnected to form a server.

3. The computing device of claim 1, wherein the internal fabric switches are configured to switch Ethernet layer 2 packets appended with routing headers.

4. The computing device of claim 1, wherein a speed of each of the plurality of fabric links is set to optimize power.

5. The computing device of claim 1, wherein the speed is dynamically adjustable based on an instantaneous utilization of a particular fabric link or an average utilization of the particular fabric link.

6. The computing device of claim 2, wherein one or more of the plurality of fabric links and the one or more Ethernet escapes each comprise a peripheral component interconnect express (PCIe) connector that is connected to the set of server boards.

7. The computing device of claim 1, further comprising a passive backplane that comprises a point-to-point server interconnect fabric.

8. The computing device of claim 1, wherein the server fabric has a topology of a tree, and wherein the one or more Ethernet escapes are disposed at each level of the tree.

9. The computing device of claim 1, wherein each of the one or more Ethernet escapes is configured to be enabled or disabled in accordance with a bandwidth that is configured to match with optimized power usage.

10. The computing device of claim 1, wherein data between the plurality of server nodes is configured to route in the server fabric but not the one or more Ethernet escapes.

11. The computing device of claim 3, wherein each of the plurality of server nodes is configured to have computational components thereon turned off in response to power reducing.

12. The computing device of claim 1, further comprising a plurality of server boards that are configured to form a rack or a backplane.

13. The computing device of claim 12, further comprising a plurality of shelves that are configured to compose a rack.

24. The computing device of claim 1, wherein each server node of the plurality of server nodes is positioned at an angled alignment relative to the server board.

14. A computing device, comprising:

a storage device of a form factor; and

a server node of the same form factor as the storage device, wherein the server node is connected to other server nodes by a plurality of fabric links, wherein a speed of each of the plurality of fabric links is configured to dynamically adjustable based on an utilization of a particular fabric link to optimize power, and wherein the server node includes;

a processor;

a memory;

an input/output circuit;

an internal fabric switch, wherein the internal fabric switch is located on the server node and is configured to connect the server node to a network and to switch information in the network; and

one or more serial advanced technology attachment (SATA) interfaces configured to connect to the storage device;

wherein the server node is configured to control the storage device.

15. The computing device of claim 14, further comprising an array of storage devices and an array of server nodes, wherein the array of server nodes are interconnected to each other.

16. The computing device of claim 14, wherein the server node is disposed within the storage device.

17. The computing device of claim 14, wherein the server node is connected to the storage device, and wherein the storage device is a local storage for the server node.

18. The computing device of claim 14, further comprising a plurality of storage devices, wherein each of the plurality of storage devices is connected to one of the one or more SATA interfaces, and wherein the server node is further configured to control the plurality of storage devices.

19. The computing device of claim 14, further comprising a plurality of server nodes and a plurality of storage devices, wherein the plurality of server nodes are interconnected to form a server fabric, and wherein the plurality of server nodes are configured to control the plurality of storage devices.

20. The computing device of claim 14, further comprising one or more Ethernet escapes and a link, wherein the one or more Ethernet escapes and the link each comprise a peripheral component interconnect express (PCIe) connector.

25. The computing device of claim 14, further comprising a server board, wherein each server node is positioned on the server board at an angled alignment relative to the server board.

21. A method for producing a high density computing system, the method comprising:

encapsulating a server node into a form factor of a disk drive and positioning the server node at an angled alignment to facilitate cooling, wherein the server node is connected to other server nodes by a plurality of fabric links, wherein a speed of each of the plurality of fabric links is configured to be dynamically adjustable based on an utilization of a particular fabric link to optimize power, and wherein the server node includes;

a processor;

a memory;

an input/output circuit;

an internal fabric switch, wherein the internal fabric switch is located on the server node and is configured to connect the server node to a network and to switch information in the network; and

one or more serial advanced technology attachment (SATA) interfaces connected to the disk drive; and

controlling, by the server node, the disk drive.

22. The method of claim 21, further comprising creating at least one internal switching pathway on a backplane of the disk drive.

23. A method for producing a high density computing system, the method comprising:

integrating a server node into a disk drive of a standard form factor and positioning the server node at an angled alignment relative to the disk drive to facilitate cooling, wherein the server node is connected to other server nodes by a plurality of fabric links, wherein a speed of each of the plurality of fabric links is configured to be dynamically adjustable based on an utilization of a particular fabric link to optimize power, and

wherein the server node includes;

a processor;

a memory;

an input/output circuit;

an internal fabric switch, wherein the internal fabric switch is located on the server node and is configured to connect the server node to a network and to switch information in the network; and

one or more serial advanced technology attachment (SATA) interfaces connected to the disk drive; and

controlling, by the server node, the disk drive.