Techniques for storing data

A method of providing access to data, comprising:

• under control of one or more computer systems configured with executable instructions,
• storing a first portion of data among a plurality of data stores, the plurality of data stores utilizing logical schemas to organize the first portion of the data and at least two of the plurality of data stores differing in at least one ability to perform a data operation;
• storing a second portion of the data in an indexed data store that is logically separate from the plurality of data stores;
• associating, at least by a mapping from which locations in the logical schemas of the data can be ascertained, data files with locations of corresponding data among the data stores, the data files represented based at least in part on a markup language;
• redistributing at least a subset of the first portion or the second portion of the data among the plurality of data stores according to at least one policy that defines, based at least in part on the at least one ability to perform a data operation, where among the plurality of data stores said at least a subset of the first portion or the second portion of the data should be stored;
• generating events indicative of changes in data contained in the plurality of data stores corresponding to said at least a subset of the first portion or the second portion of the data;
• updating the second portion of the data in the indexed data store according to the generated events;
• receiving a request from a requestor to access particular data corresponding to a data file of the data files;
• locating, based at least in part on the mapping, the particular data from the plurality of data stores;
• and
• providing the particular data to the requestor.

A system for providing access to data, comprising:

• a plurality of data stores having a first portion of the data distributed therein, the plurality of data stores utilizing logical schemas to organize the data and at least two of the plurality of data stores differing in at least one ability to perform a data operation;
• an indexed data store that is logically separate from the plurality of data stores, having a second portion of the data distributed therein;
• a data manager operable to:
• associate, at least by a mapping from which locations in the logical schemas of the data can be ascertained, data files with locations of corresponding data among the plurality of data stores, the data files represented based at least in part on a markup language;
• receive events indicative of changes in a storage location of said at least a subset of the first portion or the second portion of the data;
• redistribute said at least a subset of the first portion or the second portion of the data among the plurality of data stores or indexed data store according to at least one policy that is based at least in part on the at least one ability to perform a data operation;
• update the second portion of the data in the indexed data store according to the received events;
• receive a request from a requestor to access particular data associated with a data file of the data files;
• locate, based at least in part on the mapping, from the plurality of data stores responsive data that is associated with the data file;
• and
• provide the responsive data to the requestor.

A non-transitory computer-readable storage medium, having stored thereon instructions for causing at least one computer system to perform a method of providing access to data, the method comprising:

• storing a first portion of data among a plurality of data stores, the plurality of data stores utilizing logical schemas to organize the first portion of the data and at least two of the plurality of data stores differing in at least one ability to perform a data operation;
• storing a second portion of the data in an indexed data store that is logically separate from the plurality of data stores;
• associating, at least by a mapping from which locations in the logical schemas of the data can be ascertained, data files with locations of corresponding data among the plurality of data stores, the data files represented based at least in part on a markup language;
• redistributing at least a subset of the first portion or the second portion of the data among the plurality of data stores according to at least one policy that defines, based at least in part on the at least one ability to perform a data operation, where among the plurality of data stores said at least a subset of the first portion or the second portion of the data should be stored;
• generating events indicative of changes in data contained in the plurality of data stores corresponding to said at least a subset of the first portion or the second portion of the data;
• updating the second portion of the data in the indexed data store according to the generated events;
• receiving a request from a requestor to access particular data corresponding a data file of the data files;
• locating, based at least in part on the mapping, the particular data from the data stores;
• and
• providing the particular data to the requestor.

A computer-readable storage medium having stored thereon instructions for causing one or more computer systems to perform a method of providing access to data, the method comprising:

• receiving a request to set at least one element of a data file to at least one corresponding value, the data file formatted in a markup language;
• and
• responsive to the request,
• storing said at least one corresponding value in at least one of a plurality of data stores that collectively store a set of data, wherein at least two of the plurality of data stores differ in at least one ability to perform a data operation, wherein at least one of the plurality of data stores utilizes logical schemas to organize the data, and wherein the at least one element of the data file is associated with records of at least one of the logical schemas;
• and
• moving said at least one corresponding value from said at least one of the plurality of data stores to an indexed data store such that said at least one corresponding value is accessible by location-independent requests, the moving of said at least one corresponding value according to at least one policy that defines, based at least in part on an ability to perform a data operation, whether to store said at least one corresponding value in the plurality of data stores or the indexed data store.