Latency-aware live migration for multitenant database platforms

• US 8,880,477 B2
• Filed: 10/04/2012
• Issued: 11/04/2014
• Est. Priority Date: 10/04/2011
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Laboratories America Inc (NEC Corporation)
• Inventors: Barker, Sean, Chi, Yun, Moon, Hyun Jin, Hacigumus, Vahit Hakan
• Primary Examiner(s): MORRISON, JAY A
• Application Number:
  US13/645,103
• Publication Number:
  US 20130085998A1
• Time in Patent Office: 761 Days
• Field of Search: None
• US Class Current: 707/649
• CPC Class Codes: G06F 16/214 Database migration support
• CPC Class Codes: G06F 9/455 Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
• CPC Class Codes: G06F 9/5088 involving task migration

Abstract

Methods and systems for database migration from a multitenant database include taking a snapshot of an original database to be migrated with a hot backup process, such that the database is still capable of answering queries during the hot backup process; maintaining a query log of all queries to the tenant database after the hot backup process begins; initializing a new database at a target server using the snapshot; replaying the query log synchronize the new database with the original database; and answering new queries with the new database and not the original database.

Claims

1. A method for migration from a multitenant database, comprising:

taking a snapshot of an original database to be migrated with a hot backup process, such that the database is still capable of answering queries during the hot backup process;

selecting a database migration method that does not generate a service level agreement (SLA) violation and otherwise selecting a method that generates the fewest SLA violations during the database migration;

applying a proportional-integral-derivative (PID) controller to determine a speed of migration based on system performance, wherein the PID controller has a proportional path, an integral path and a derivative path, each path is scaled by coefficients K_p, K_i, and K_d respectively, wherein an output at time t with error e(t) is given by;

2. The method of claim 1, further comprising computing a target level of performance based on meeting service level agreement (SLA) guarantees.

3. The method of claim 2, further comprising moderating a speed of the hot backup process based on current and historic system performance.

4. The method of claim 3, wherein moderating the speed comprises comparing current and historic system performance to the computed target level of performance to whether to increase or decrease speed in accordance with a respective shortfall or surplus relative to the target level of performance.

5. The method of claim 3, wherein moderating the speed comprises allocating resources to the hot backup process to optimize a backup speed while maintaining SLA guarantees.

6. The method of claim 5, wherein the resources comprise a storage input/output rate and a network bandwidth.

7. The method of claim 3, wherein moderating the speed comprises using a proportional-integral-derivative controller to adaptively respond to changes in an amount of slack resources.

8. The method of claim 1, wherein the amount of slack resources is determined by comparing the measured transaction latency to a target transaction latency that represents an optimal balance between speed of the backup process and transactions for other tenants.

9. A multitenant database system, comprising:

code to select a database migration method that does not generate a service level agreement (SLA) violation and otherwise select a method that generates the fewest SLA violations during the database migration;

a proportional-integral-derivative (PID) controller coupled to the database system to determine a speed of migration based on system performance, wherein the PID controller has a proportional path, an integral path and a derivative path, each path is scaled by coefficients K_p, K_i, and K_d respectively, wherein an output at time t with error e(t) is given by;