Devices and methods for interconnecting server nodes

Patent metadata

Field	Value
Publication number	US-2016004445-A1
Application number	US-201514852812-A
Country	US
Kind code	A1
Filing date	Sep 14, 2015
Priority date	Jun 19, 2012
Publication date	Jan 7, 2016
Grant date	—

Abstract

Described are aggregation devices and methods for interconnecting server nodes. The aggregation device can include an input region, an output region, and a memory switch. The input region includes a plurality of input ports. The memory switch has a shared through silicon via (TSV) memory coupled to the input ports for temporarily storing data received at the input ports from a plurality of source devices. The output region includes a plurality of output ports coupled to the TSV memory. The output ports provide the data to a plurality of destination devices. A memory allocation system coordinates a transfer of the data from the source devices to the TSV memory. The output ports receive and process the data from the TSV memory independently of a communication from the input ports.

First claim

What is claimed is:

1 . An aggregation device, comprising:
a plurality of input ports; a plurality of output ports; a memory coupled between the input ports and the output ports; and a memory allocator, the memory allocator coordinating transfers of data from the input ports to the memory and from the memory to the output ports, the transfers of data from the input ports to the memory being independent of the transfers of data from the memory to the output ports.

2 . The aggregation device of claim 1 , further comprising:
a separate memory output queue associated with each output port; wherein, for each unit of data in the memory that is to be transferred to an output port, a record of the unit of data is stored in the memory output queue associated with the output port, the record comprising an indication of a location of the unit of data in the memory.

3 . The aggregation device of claim 2 , wherein, when the memory allocator transfers a given unit of data from an input port to the memory, the memory allocator:
generates a given record based on information associated with the given unit of data; stores the given record in a memory output queue associated with an output port to which the given unit of data is to be transferred from the memory; and stores the given unit of data in the memory.

4 . The aggregation device of claim 3 , wherein, when the memory allocator transfers the given unit of data from the memory to an output port, the memory allocator:
reads the given record for the given unit of data from the corresponding memory output queue; acquires the given unit of data from the memory based on the indication of the location of the given unit of data in the memory from the given record; and transfers the given unit of data to the corresponding output port.

5 . The aggregation device of claim 3 , wherein the given unit of data is stored in the memory and the record for the given unit of data is separately stored in the memory output queue.

6 . The aggregation device of claim 2 , wherein the record for each unit of data comprises state data for maintaining legacy protocol transparency.

7 . The aggregation device of claim 1 , wherein the memory allocator comprises a buffer descriptor manager, the buffer descriptor manager allocating buffers in the memory for storing data transferred to the memory from the input ports.

8 . The aggregation device of claim 7 , wherein the buffer descriptor manager performs wear leveling when allocating buffers.

9 . The aggregation device of claim 1 , further comprising:
a write controller, the write controller receiving data at the input ports and transferring the data to the memory; and a read controller, the read controller acquiring data from the memory and transferring the data to the output ports.

10 . The aggregation device of claim 1 , wherein the memory comprises a plurality of memory partitions, each memory partition associated with at least some of the input ports and at least some of the output ports.

11 . The aggregation device of claim 1 , wherein the memory is a through-silicon via (TSV) memory.

12 . The aggregation device of claim 11 , wherein the TSV memory comprises a stack of two or more memory dies.

13 . A server aggregation system comprising:
one or more controllers; and a memory switch coupled to the one or more controllers, the memory switch comprising:
a plurality of input ports;
a plurality of output ports;
a memory coupled between the input ports and the output ports; and
a memory allocator, the memory allocator coordinating transfers of data from the input ports to the memory and from the memory to the output ports, the transfers of data from the input ports to the memory being independent of the transfers of data from the memory to the output ports.

14 . The server aggregation system of claim 13 , wherein the memory switch further comprises:
a separate memory output queue associated with each output port; wherein, for each unit of data in the memory that is to be transferred to an output port, a record of the unit of data is stored in the memory output queue associated with the output port, the record comprising an indication of a location of the unit of data in the memory.

15 . The server aggregation system of claim 14 , wherein, when the memory allocator transfers a given unit of data from an input port to the memory, the memory allocator:
generates a given record based on information associated with the given unit of data; stores the given record in a memory output queue associated with an output port to which the given unit of data is to be transferred from the memory; and stores the given unit of data in the memory.

16 . The server aggregation system of claim 15 , wherein, when the memory allocator transfers the given unit of data from the memory to an output port, the memory allocator:
reads the given record for the given unit of data from the corresponding memory output queue; acquires the given unit of data from the memory based on the indication of the location of the given unit of data in the memory from the given record; and transfers the given unit of data to the corresponding output port.

17 . The server aggregation system of claim 14 , wherein the record for each unit of data comprises state data for maintaining legacy protocol transparency.

18 . The server aggregation system of claim 13 , wherein the memory allocator comprises a buffer descriptor manager, the buffer descriptor manager allocating buffers in the memory for storing data transferred to the memory from the input ports, wherein the buffer descriptor manager performs wear leveling when allocating buffers.

19 . The server aggregation system of claim 13 , wherein the memory comprises a plurality of memory partitions, each memory partition associated with at least some of the input ports and at least some of the output ports.

20 . The server aggregation system of claim 13 , wherein the memory is a through-silicon via (TSV) memory.

Assignees

• Mayhew David E
• Hummel Mark D
• Osborn Michael J
• Advanced Micro Devices Inc

Inventors

• Mayhew David E
• Hummel Mark D
• Osborn Michael J

Classifications

• G06F3/0673

  Single storage device · CPC title

• G06F3/0604 Primary

  Improving or facilitating administration, e.g. storage management · CPC title

• G06F3/0631

  by allocating resources to storage systems · CPC title

• H04L49/103 Primary

  using a shared central buffer; using a shared memory · CPC title