Robust data replication among data centers

Patent metadata

Field	Value
Publication number	US-11314772-B2
Application number	US-201716605924-A
Country	US
Kind code	B2
Filing date	Apr 26, 2017
Priority date	Apr 26, 2017
Publication date	Apr 26, 2022
Grant date	Apr 26, 2022

Abstract

A method of data replication between a first data center and a second data center, wherein the first data center has a data set to be replicated, the method comprising: determining (501) whether a first path from the first data center to the second data center is suitable for the data set to be replicated, by: transmitting (502) test data along the first path, taking (503) latency measurements for the test data along the first path; and determining a latency parameter of the first path using the latency measurements; and determining (504) if the latency parameter of the test data is less than a maximum latency threshold. If the latency parameter is less than the maximum latency threshold, the method performs (505) replication of the data set using the first path.

First claim

The invention claimed is:

1. A method of data replication between a first data center and a second data center, wherein the first data center has a data set to be replicated, the method comprising:
determining whether a first path from the first data center to the second data center is suitable for the data set to be replicated, by:
transmitting test data along the first path;
taking latency measurements for the test data along the first path;
determining a latency parameter of the first path using the latency measurements; and
determining if the latency parameter of the test data is less than a maximum latency threshold; and

performing or not performing replication of the data set using the first path, depending respectively on whether or not the latency parameter is determined to be less than the maximum latency threshold.

2. A method of data replication according to claim 1 , wherein the latency parameter of the first path is determined to be greater than the maximum latency threshold, and wherein the method further comprises:
identifying a second path between the first data center and the second data center;
determining whether the second path from the first data center to the second data center is suitable for the data set to be replicated, by:
transmitting test data along the second path;
taking latency measurements for the test data along the second path;
determining a latency parameter of the second path using the latency measurements taken for the test data along the second path; and
determining if the latency parameter of the second path is less than the maximum latency threshold; and

performing or not performing replication of the data set using the second path, depending respectively on whether or not the latency parameter of the second path is determined to be less than the maximum latency threshold.

3. A method of data replication according to claim 1 , further comprising:
identifying a group of paths between the first data center and the second data center; and
if a respective latency parameter for each path in the group is greater than the maximum latency threshold, identifying that the second data center cannot be used for synchronous replication; and,
if the second data center cannot be used:
identifying a third data center which has a path to the first data center, the path having a latency parameter less than the maximum latency threshold; and
performing replication of the data set at the third data center.

4. The method according to claim 1 , wherein performing or not performing replication of the data set using the first path comprises performing replication of the data set using the first path, wherein said performing comprises, at an application running in the first data center:
receiving a replication wait time;
transmitting the data set to the second data center;
upon transmission of the data set to the second data center, suspending at least one part of a process of the application;
after elapse of a period equal to the replication wait time, resuming the at least one part of the process of the application.

5. The method according to claim 4 further comprising receiving an acknowledgement from the second data center.

6. The method according to claim 1 , wherein performing or not performing replication of the data set using the first path comprises performing replication of the data set using the first path, wherein said performing comprises, at a controller for the first data center:
receiving the data set from an application;
transmitting the data set to the second data center; and
after a replication wait period, sending an indication to the application that the data set has been replicated, wherein the replication wait period is calculated using the latency parameter of the first path.

7. The method according to claim 6 further comprising receiving an acknowledgement from the second data center.

8. The method according to claim 6 , wherein the replication wait period is a period calculated using the latency parameter of the first path and an estimated write period at the second data center.

9. The method according to claim 1 , wherein determining the latency parameter of the first path comprises:
determining a mean of the latency measurements taken for the test data along the first path;
determining a standard deviation of the latency measurements taken for the test data along the first path;
setting a required probability, the required probability being a probability that the data set will suffer a latency less than the latency parameter; and
using the mean, standard deviation and required probability to set the latency parameter.

10. The method according to claim 1 , wherein latency estimation is performed periodically during a data replication.

11. A controller for a data center network, the controller comprising:
a processor;
a memory for storing data and instructions;
a transmitter for transmitting data to the data center network; and
a receiver for receiving data from the data center network;
wherein the processor is configured to:
transmit test data along a first path from a first data center to a second data center;
take latency measurements for the test data along the first path;
determine a latency parameter of the first path using the latency measurements;
determine if the latency parameter of the test data is less than a maximum latency threshold; and
perform or not perform replication of the data set using the first path, depending respectively on whether or not the latency parameter is determined to be less than the maximum latency threshold.

12. The controller according to claim 11 , wherein the latency parameter of the first path is determined to be greater than the maximum latency threshold, and wherein the processor is further configured to:
identify a second path between the first data center and the second data center;
determine whether the second path from the first data center to the second data center is suitable for the data set to be replicated, by:
transmitting test data along the second path;
taking latency measurements for the test data along the second path;
determining a latency parameter of the second path using the latency measurements taken for the test data along the second path; and
determining if the latency parameter of the second path is less than the maximum latency threshold; and

perform or not perform replication of the data set using the second path, depending respectively on whether or not the latency parameter of the second path is determined to be less than the maximum latency threshold.

13. A controller according to claim 11 , wherein the processor is further configured to:
identify a group of paths between the first data center and the second data center; and
if a respective latency parameter for each path in the group is greater than the maximum latency threshold, identify that the second data center cannot be used for synchronous replication; and,
if the second data center cannot be used:
identify a third data center which has a path to the first data center, the path having a latency parameter less than the maximum latency threshold; and
perform replication of the data set at the third data center.

14. A controller according to claim 11 , wherein the processor is configured to perform replication of the data set using the first path by, at an application running in the first data center:
receiving a replication wait time;
transmitting the data set to the second data center;
upon transmission of the data set to the second data center, suspending at least one part of a

Assignees

• Ericsson Telefon Ab L M

Inventors

• Bottari Giulio
• Ruffini Stefano

Classifications

• G06F16/27 Primary

  Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor · CPC title

• G06F11/3495

  for systems · CPC title

• G06F11/3006

  where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems (multiprogramming arrangements G06F9/46; allocation of resources G06F9/50) · CPC title

• H04L43/16

  Threshold monitoring · CPC title

• G06F11/2007

  using redundant communication media · CPC title