Browse Prior Art Database

Peer-to-Peer Remote Copy Relationship Organization

IP.com Disclosure Number: IPCOM000243107D
Publication Date: 2015-Sep-15
Document File: 3 page(s) / 40K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed are a method and system for allowing a peer-to-peer relationship structure to be used as either a source or target. Many of the fields are common, such as peer device identification information, while unique fields are stored in a union.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 45% of the total text.

Page 01 of 3

Peer-to-Peer Remote Copy Relationship Organization

A peer-to-peer remote copy source device can have multiple targets. A peer-to-peer remote copy target device can have only one source. A device structure contains an array of source relationship structures and one target structure.

When a peer-to-peer remote copy device relationship is established, a source relationship structure is allocated in the source system and populated with information describing its target, and a target relationship structure is allocated in the target system to store information that describes the source.

On a failover, one of a device's source relationship structures is allocated and populated with peer identifying information from the target relationship structure. The target relationship structure is then freed. Optionally, the source structure can also be populated with information describing a peer other than the target's source.

On a failback, a device's target relationship structure is allocated and populated with peer descriptor information from one of its source relationship structures. The source relationship structure is then freed.

A problem with out-of-source relationships on failover can occur in the event that a device's, d 's, source structure array is
of length n, and device d is a target with an associated source, c, and a source to n targets; therefore, device d cannot
be in any more relationships. If device c tries to copy to d, it cannot because it has no targets defined; so, the user tries
to issue a failover c -d so it can access associated data at d and track d 's delta with c, so it can later resynchronize with c . The failover fails because device d has no available source relationship structures.

Another problem is inconsistency on crash in the middle of conversion. If a node crashes in the middle of a failover or failback, then a relationship that is being converted from one type to another (e.g., source to target), then the system is subject to an inconsistency wherein both relationships exist at the same time.

In addition, for linear time lookup, to find a relationship or determine if a relationship exists, each entry in the source array is checked until one is found that matches or all entries in the array are examined.

Further, current methods allow limited relationships per device. With a fixed number of relationships per device, one device may have used all of its relationships while another has some available. A user cannot establish another

relationship unless one of the relationships is removed.

1


Page 02 of 3

The proposed solution is to allow a relationship structure to be used as either a source or target. Many of the fields are common, such as peer device identification information. Unique fields are stored in a union.

In addition, the novel solution allocates relationships from a system or node-wide pool. The approach defines a relationship key consisting of a local device plus the associated peer system and device....