HIERARCHICAL SYNCHRONIZATION OF REPLICAS

• US 20090006489A1
• Filed: 06/29/2007
• Published: 01/01/2009
• Est. Priority Date: 06/29/2007
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Rodeheffer, Thomas L., Terry, Douglas B., Wobber, Edward P., Ramasubramanian, Venugopalan Saraswati
• Application Number:
  US11/771,273
• Publication Number:
  US 20090006489A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/1787 Details of non-transparently synchronising file systems
• CPC Class Codes: G06F 16/184 implemented as replicated file system
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: H04L 67/1095 Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes

Abstract

A hierarchical system is disclosed for synchronizing partially-replicated collections that provides guaranteed paths of information to all replicas in a collection. Each partial replica is assigned a parent replica to act as a proxy on its behalf, and with which the replica synchronizes regularly. Each parent replica is responsible for one or more child replicas. Parent replicas have filters which are either the same as, or more inclusive than, their child replicas, and parent replicas thus store and synchronize all objects that are of interest to their one or more child replicas.

Claims

1. A computer implemented method of synchronizing a plurality of replicas having a plurality of items in a weakly consistent distributed collection, the method comprising the steps of:

(a) establishing a first partial replica in the weakly consistent distributed collection, the first partial replica having filter defining a first interest set of items including less than all of the items in the plurality of items;

(b) establishing a second partial replica in the weakly consistent distributed collection, the second partial replica having filter defining second interest set of items including less than all of the items in the plurality of items, the first interest set of the first replica and the second interest set of the second replica sharing at least one item in common from the plurality of items; and

(c) guaranteeing a synchronization path between the first and second replicas.

2. A computer implemented method as recited in claim 1, further comprising the steps of:

(d) establishing one or more parent replicas in the weakly consistent distributed collection, the one or more parent replicas each having a filter defining an interest set of items that is the same as or less restrictive than the filter of the first partial replica; and

(e) guaranteeing a synchronization path between the first partial replica and the one or more parent replicas.

3. A computer implemented method as recited in claim 2, wherein said step (d) of establishing one or more parent replicas in the weakly consistent distributed collection comprises the step of establishing multiple parent replicas.

4. A computer implemented method as recited in claim 3, wherein the filter of the first partial replica is a split filter divided into first and second sub-filters, the first partial replica synchronizing with a first parent replica of the multiple parent replicas for items in the interest set of the first sub-filter, and the first partial replica synchronizing with a second parent replica of the multiple parent replicas for items in the interest set of the second sub-filter.

5. A computer implemented method as recited in claim 1, further comprising the step (e) of maintaining a push-out store in the first partial replica, the push-out store maintaining an item that has changed from within the first interest set to be outside of the first interest set.

6. A computer implemented method as recited in claim 5, the weakly consistent distributed collection further including a parent replica having a synchronization path with the first partial replica, and having a filter defining an item set that is the same as or less restrictive than the first partial replica, the method further comprising the step (f) of the first partial replica notifying the parent replica of the item in the push-out store of the first partial replica.

7. A computer implemented method as recited in claim 6, the weakly consistent distributed collection further including a grandparent replica having a synchronization path with the parent replica, and having a filter defining an item set that is the same as or less restrictive than the parent replica, the method further comprising the step (g) of the parent replica notifying the grandparent replica of the push-out notification the parent received in said step (f).

8. A computer implemented method as recited in claim 6, the weakly consistent distributed collection further including a reference replica having an interest set including all items in the plurality of items and a synchronization path to the first partial replica by one or more intermediate replicas including the parent replica, the method further comprising the step (h) of the reference replica receiving notification of the item in the push-out store of the first partial replica via the one or more intermediate replicas.

9. A computer implemented method as recited in claim 6, the weakly consistent distributed collection further including a reference replica having an interest set including all items in the plurality of items, the method further comprising the step (i) of the first partial replica notifying a given replica in the collection that is not a parent replica of the first partial replica, the given replica having a synchronization path to the reference replica.

10. A computer implemented method as recited in claim 5, the weakly consistent distributed collection further including a parent replica having a synchronization path with the first partial replica, and having a filter defining an item set that is the same as or less restrictive than the first partial replica, the method further comprising the steps of:

(j) the first partial replica determining whether the item in the push-out store is within the interest set of the parent replica; and

(k) the first partial replica sending an identification and version of the item in the push-out store if it is determined in said step (j) that the item in the push-out store is within the interest set of the parent replica.

11. A computer implemented method as recited in claim 5, wherein said step (e) of maintaining an item in the push-out store comprises the steps:

(e1) setting a responsibility bit for an item when the item is created or updated by the first partial replica; and

(e2) storing the item in the push-out store when the item falls outside of the first partial replica'"'"'s interest set only if the responsibility bit for the item is set in said step (e1).

12. A computer implemented method as recited in claim 11, the weakly consistent distributed collection further including a parent replica having a synchronization path with the first partial replica, and having a filter defining an item set that is the same as or less restrictive than the first partial replica, the method further comprising the step (f) of the first partial replica notifying the parent replica of the item in the push-out store of the first partial replica only if the responsibility bit for the item is set in said step (e1).

13. A computer implemented method of synchronizing a plurality of replicas having a plurality of items in a weakly consistent distributed collection, the method comprising the steps of:

(a) defining a first filter in a first replica of the plurality of replicas, the first filter defining an interest set of items for the first replica;

(b) defining a second filter in a second replica of the plurality of replicas, the second filter defining an interest set of items for the second replica and the second filter being the same as or more inclusive than the first filter;

(c) guaranteeing a synchronization path between the first replica and the second replica;

(d) defining a third replica of the plurality of replicas to have knowledge of all items in the collection;

(e) requiring a synchronization path between the second replica and the third replica; and

(f) allowing ad hoc synchronization between the plurality of replicas in the weakly consistent distributed collection.

14. A computer implemented method as recited in claim 13, wherein the weakly consistent distributed collection includes a fourth replica having a synchronization path with the first replica, and having a filter being the same as or more inclusive than the first filter, and wherein the first filter of the first replica is a split filter divided into first and second sub-filters, the first replica synchronizing with the second replica for items in the interest set of the first sub-filter, and the first replica synchronizing with a fourth replica for items in the interest set of the second sub-filter.

15. A computer implemented method as recited in claim 13, further comprising the step (g) of maintaining a push-out store in the first replica, the push-out store maintaining an item that has changed from within the first interest set to be outside of the first interest set.

16. A computer implemented method as recited in claim 15, further comprising the step (h) of the first replica notifying the second replica of the item in the push-out store of the first replica.

17. A collection of replicas having knowledge of items, the replicas communicating with each other in a weakly-consistent distributed ad hoc network, the collection of replicas comprising:

a reference replica having an interest set including all items in the collection;

a first replica having a filter indicating an interest set of items the first replica receives;

a proxy replica having a filter defined to include at least a portion of the filter of the first replica, a guaranteed synchronization pathway existing between the first replica and the proxy replica and between the proxy replica and the reference replica allowing knowledge of the first replica to be shared with the reference replica and allowing knowledge of the reference replica to be shared with the first replica.

18. A collection of replicas as recited in claim 17, wherein the plurality of replicas ensure that a modification to an item is passed from the modifying replica to the reference replica via a synchronization path including one or more intermediate replicas regardless of whether the modified item is inside or outside of the filters of the intermediate replicas.

19. A collection of replicas as recited in claim 17, wherein the proxy replica comprises a first proxy replica, the collection of replicas further comprising a second proxy replica having a filter defined to include at least a portion of the filter of the first replica, a guaranteed synchronization pathway existing between the first replica and the second proxy replica and between the second proxy replica and the reference replica.

20. A collection of replicas as recited in claim 18, the filter in the first replica comprising a first sub-filter and a second sub-filter, the first proxy replica including the first sub-filter and the second proxy replica including the second sub-filter.