Data storage system with redundant internal networks

Patent metadata

Field	Value
Publication number	US-11438411-B2
Application number	US-202017013430-A
Country	US
Kind code	B2
Filing date	Sep 4, 2020
Priority date	Dec 28, 2016
Publication date	Sep 6, 2022
Grant date	Sep 6, 2022

Abstract

A data storage system includes a rack, multiple head nodes, multiple data storage sleds, and at least two networking devices. The at least two network devices are configured to implement at least two redundant networks within the data storage system. Also, each of the head nodes is assigned at least two network addresses for communication with the data storage sleds of the data storage system via the at least two networking devices. The data storage sleds each include multiple mass storage devices and a sled controller that is configured to couple with the at least two network switches. IO In some embodiments, the data storage system further includes redundant power systems within a rack in which the head nodes, the data storage sleds, and the at least two networking devices are mounted.

First claim

What is claimed is:

1. A data storage system, comprising:
a rack;
a plurality of head nodes mounted in the rack;
a plurality of data storage sleds mounted in the rack; and
at least two networking devices mounted in the rack;
wherein the at least two networking devices are configured to provide at least two redundant networks in the rack between the plurality of head nodes and the plurality of data storage sleds, such that during a failure of one of the at least two redundant networks in the rack, each of the plurality of head nodes in the rack can communicate with each of the plurality of data storage sleds in the rack through another one of the at least two redundant networks; and
wherein to provide the at least two redundant networks:
the each of the plurality of head nodes is coupled to the each of the plurality of data storage sleds via a first one of the at least two networking devices;
the each of the plurality of head nodes is also coupled to the each of the plurality of data storage sleds via a second one of the at least two network devices; and
the each of the plurality of head nodes is assigned at least two private IP addresses for communicating via respective ones of the at least two redundant networks with the data storage sleds and at least one additional public IP address for communicating with client devices external to the data storage sleds,
wherein for a given one of the plurality of head nodes, a first one of the at least two private IP address is used for routing traffic between the given head node and the data storage sleds via a first one of the at least two redundant networks, and a second one of the at least two private IP address is used for routing traffic between the given head node and the data storage sleds via a second one of the at least two redundant networks.

2. The data storage system of claim 1 , wherein each of the head nodes comprises:
at least two redundant network interface cards, wherein:
a first one of the at least two redundant network interface cards couples the respective head node to a first one of the at least two redundant networks; and
a second one of the at least two redundant network interface cards couples the respective head node to a second one of the at least two redundant networks.

3. The data storage system of claim 1 , wherein each of the head nodes comprises:
a set of at least two ports configured to connect the respective head node to each of the at least two redundant networks; and
an additional set of two or more ports configured to couple the respective head node to one or more clients external to the data storage system.

4. The data storage system of claim 1 , wherein each of the head nodes is:
further assigned at least another public IP address for communicating with the client devices external to the data storage system.

5. The data storage system of claim 1 , wherein each of the at least two redundant networks connects respective ones of the head nodes to respective ones of the data storage sleds via a single network hop within the rack.

6. The data storage system of claim 1 , wherein each of the at least two redundant networks connects respective ones of the head nodes to other respective ones of the head nodes via a single network hop within the rack.

7. The data storage system of claim 1 , further comprising:
at least two connectors for coupling the data storage system to at least two independent electrical power sources; and
at least two power distribution systems within the rack configured to supply redundant power from the at least two independent electrical power sources to each of the head nodes and to each of the data storage sleds.

8. The data storage system of claim 1 , wherein each of the head nodes is configured to:
receive a request to write data to a volume stored in the data storage system;
write the data included with the request to a storage drive of the respective head node; and
cause the data to be replicated to another respective head node in the rack via one of the at least two redundant networks in the rack.

9. The data storage system of claim 8 , wherein if a connection to the other respective head node is not available via the one of the at least two redundant networks, the respective head node is configured to cause the data to be replicated to the other respective head node via another one of the at least two redundant networks in the rack.

10. The data storage system of claim 1 , wherein each of the head nodes is configured to:
flush data stored in a storage of the respective head node to a sub-set of the plurality of data storage sleds via one of the at least two redundant network in the rack,
wherein respective portions of the data transmitted to the sub-set of the plurality of data storage sleds are stored in respective portions of storage devices included in the sub-set of the plurality of data storage sleds.

11. The data storage system of claim 10 , wherein if a connection to the sub-set of the data storage sleds is not available via the one of the at least two redundant networks, the respective head node is configured to flush the data to the sub-set of the data storage sleds via another one of the at least two redundant networks in the rack.

12. The data storage system of claim 1 , wherein at least one of the head nodes comprises:
a log-based storage configured to store write data for a volume stored in the data storage system; and
at least two redundant network interface cards,
wherein respective ones of the at least two redundant network interface cards are configured to couple the head node to respective ones of the at least two networking devices mounted in a same rack with the head node,
wherein the at least one head node is configured to:
communicate with other head nodes mounted in the same rack with the head node via redundant networks implemented in the same rack with the head node that are implemented via the at least two networking devices coupled to the network interface cards of the head node; and
communicative with data storage sleds mounted in the same rack with the head node via the redundant networks implemented in the same rack with the head node via the at least two networking devices coupled to the network interface cards of the head node.

13. The data storage system of claim 12 , wherein the at least one head node is configured to use the at least two private IP addresses for communicating with the other head nodes and data storage sleds in the same rack with the head node via the redundant networks implemented in the same rack with the head node.

14. The data storage system of claim 13 , wherein the head node is assigned at least another public IP address for communicating with the client devices external to the rack.

15. The data storage system of claim 1 , wherein at least one of the data storage sleds comprises:
a plurality of mass storage devices; and
a storage controller comprising at least two network interfaces configured to couple the at least one data storage sled to the plurality of head nodes mounted in a same rack with the at least one data storage sled via the at least two redundant networks implemented in the rack.

16. The data storage system of claim 15 , wherein:
the at least one data storage sled is assigned at least two network addresses for communicating with respective ones of the head nodes via the at least two networking devices mounted in the rack with the data storage sleds and the head nodes.

17. The data storage system of claim 16 , wherein the at least two network addresses assigned

Assignees

• Amazon Tech Inc

Inventors

• Kusters Norbert P
• Arumugam Nachiappan
• Watson Christopher Nathan
• Brooker Marc John
• Richardson David R
• Wei Danny
• Guthrie Ii John Luther

Classifications

• G06F11/2015

  Redundant power supplies (power supply failure G06F1/30) · CPC title

• G06F3/0647

  Migration mechanisms · CPC title

• G06F3/065

  Replication mechanisms · CPC title

• G06F3/0685

  Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays · CPC title

• H04L67/568

  Storing data temporarily at an intermediate stage, e.g. caching · CPC title