Code-division-multiple-access (CDMA)-based network coding for massive memory servers

• US 9,930,114 B2
• Filed: 10/28/2014
• Issued: 03/27/2018
• Est. Priority Date: 10/28/2014
• Status: Active Grant

• Current Assignee: Empire Technology Development LLC (Allied Inventors Management, LLC)
• Original Assignee: Empire Technology Development LLC (Allied Inventors Management, LLC)
• Inventors: Ma, Xudong
• Primary Examiner(s): Keller, Michael A
• Application Number:
  US14/526,434
• Publication Number:
  US 20160119423A1
• Time in Patent Office: 1,246 Days
• CPC Class Codes: H04B 1/707 using direct sequence modulation
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]

Abstract

Technologies are generally described for code-division-multiple-access (CDMA)-based network-coding for reading data from massive memory servers. According to some examples, data may be modulated by spreading sequences prior to being transmitted from one memory node to another. In addition, the received signals (modulated data) from multiple memory nodes may be combined by a receiver memory node before being forwarded to other memory nodes arranged in a grid of memory nodes. The memory nodes may be assigned communication bandwidths flexibly and rapidly by changing the respective spreading sequences, which may be orthogonal or near-orthogonal for different memory nodes to allow support of random-access burst or dynamic data traffic, and enhancing fault-tolerance.

Claims

1. A method for code-division-multiple-access (CDMA)-based network-coding for a network, the method comprising:

receiving a first signal at a first memory node from a second memory node, wherein the first signal comprises data modulated with a first spreading sequence that is dynamically assigned to the first signal;

receiving a second signal at the first memory node from a third memory node, wherein the second signal comprises data modulated with a second spreading sequence that is dynamically assigned to the second signal, and

wherein the dynamic assignment of the first spreading sequence and the second spreading sequence is based on one or more network parameters of the network;

combining the received first signal with the received second to form a combined signal at the first memory node;

transmitting the combined signal to a fourth memory node that neighbors the first memory node, wherein the first memory node, the second memory node, the third memory node, and the fourth memory node comprise data storage devices within a massive memory server and are communicatively coupled to neighboring memory nodes;

detecting a failure of at least one of the second memory node and the third memory node; and

transmitting a constant data stream via the first memory node to the fourth memory node without the transmission of the combined signal being affected.

2. The method of claim 1, wherein combining includes:

combining the received first signal and the received second signal, at the first memory node, with a third signal received at the first memory node from, a fifth memory node to form the combined signal, wherein the third signal comprises data modulated with a third spreading sequence by the fifth memory node.

3. The method of claim 2, further comprising:

generating a fourth signal at the first memory node by modulating data with a fourth spreading sequence; and

transmitting the fourth signal to the fourth memory node that neighbors the first memory node.

4. The method of claim 3, further comprising:

selecting the fourth spreading sequence based on a communication bandwidth usable for the first memory node, wherein the first spreading sequence, the second spreading sequence, and the fourth spreading sequence are distinct from each other.

5. The method of claim 4, wherein selecting the fourth spreading sequence comprises one of:

selecting a longer spreading sequence in response to a determination that the first memory node is assigned a lower communication rate compared to other memory nodes;

selecting a shorter spreading sequence in response to a determination that the first memory node is assigned a higher communication rate compared to other memory nodes;

or

selecting the fourth spreading sequence to be an orthogonal spreading sequence relative to the second spreading sequence and the third spreading sequence.

6. The method of claim 1, further comprising:

transmitting a specific signal to the fourth memory node after one of the second memory node and the third memory node is determined to have failed.

7. The method of claim 1, further comprising:

multiplying each received signal with an amplification coefficient at the first memory node.

8. The method of claim 1, wherein combining the received first signal with the received second signal comprises combining the received first signal and the received second signal by employing one of an analog summation or a digital summation.

9. A massive memory server configured to employ code-division-multiple-access (CDMA)-based network-coding for a network, the server comprising:

one or more processors; and

a plurality of interconnected memory nodes communicatively coupled to the one or more processors, wherein at least one memory node of the plurality of interconnected memory nodes is configured to be controlled by the one or more processors to;

dynamically assign, a respective spreading sequence to each respective memory node, based on one or more network parameters of the network;

modulate data stored at each respective memory node with the assigned respective spreading sequence to generate a signal;

transmit the signal to at least one neighbor memory node;

receive at least two other signals from at least two other neighbor memory nodes;

combine the at least two other signals to form a combined signal;

transmit the combined signal from the at least one memory node to the at least one neighbor memory node;

detect a failure of one of the at least two other neighbor memory nodes; and

transmit a constant data stream via the failed one of the at least two other neighbor memory nodes to the at least one neighbor memory node without the transmission of the combined signal being affected.

10. The server of claim 9, wherein the plurality of interconnected memory nodes is coupled by interconnection networks organized into a grid, and wherein each memory node of the plurality of interconnected memory nodes is coupled to neighbor memory nodes through one of a wired connection or an optical connection.

11. The server of claim 9, wherein each memory node of the plurality interconnected memory nodes is further configured to translate the data into a binary representation prior to modulation.

12. The server of claim 9, wherein each memory node of the plurality interconnected memory nodes is further configured to employ a random-access, burst memory-access communication protocol.

13. The server of claim 9, wherein each memory node of the plurality of interconnected memory nodes includes one of a flash memory device, a dynamic random access memory (DRAM) device, or a static random access memory (SRAM) device.

14. The server of claim 9, wherein the one or more processors are configured to:

select a different spreading sequence for each memory node of the plurality of interconnected memory nodes based on a communication bandwidth assigned to each memory node; and

assign different communication bandwidths dynamically to different memory nodes of the plurality of interconnected memory nodes by change of respective spreading sequences.

15. A system configured to read from massive memory servers through code-division-multiple-access (CDMA)-based network-coding for a network, the system comprising:

one or more massive memory servers associated with a network of memory nodes that are organized into a grid, wherein each memory node is wirelessly or wiredly coupled to neighbor memory nodes, and

wherein at least one memory node of the network of memory nodes is configured to;

dynamically assign a respective spreading sequence to each respective memory node based on one or more network parameters of the network;

modulate data stored at each respective memory node with the assigned respective spreading sequence to generate a signal;

transmit the signal to at least one neighbor memory node;

receive at least two other signals from at least two other neighbor memory nodes;

combine the at least two other signals to form a combined signal;

transmit the combined signal from the at least one memory node to the at least one neighbor memory node;

detect a failure of one of the at least two other neighbor memory nodes; and

transmit a constant data stream via the failed one of the at least two other neighbor memory nodes to the at least one neighbor memory node without the transmission of the combined signal being affected.

16. The system of claim 15, wherein the assigned respective spreading sequences for each respective memory node includes orthogonal spreading sequences.

17. The system of claim 16, wherein the massive memory servers are configured to enable multiple communication flows to occupy a same transmission line within networks of memory nodes without explicit resource allocation through the orthogonal spreading sequences.

18. The system of claim 15, wherein the massive memory servers are configured to select:

a longer spreading sequence for memory nodes that are assigned a lower communication rate compared to other memory nodes; and

a shorter spreading sequence for memory nodes that are assigned a higher communication rate compared to other memory nodes.

19. The system of claim 15, wherein the at least one memory node of the one or more massive memory servers is included in a single massive memory server such that the data modulation, the signal transmission, the reception of the at least two other signals, the combination of the at least two other signals, and the transmission of the combined signal are performed in the single massive memory server.

20. The system of claim 15, wherein the at least one memory node of the one or more massive memory servers includes a plurality of memory nodes associated with a virtual single massive memory server such that the data modulation, the signal transmission, the reception of the at least two other signals, the combination of the at least two other signals, and the transmission of the combined signal are performed in the virtual single massive memory server.

21. The system of claim 15, wherein the at least one memory node of the one or more massive memory servers includes a plurality of memory nodes associated with a plurality of the massive memory servers such that at least some of the data modulation, the signal transmission, the reception of the at least two other signals, the combination of the at least two other signals, and the transmission of the combined signal are performed in a distributed manner amongst the plurality of memory nodes associated with the plurality of massive memory servers.