Method and apparatus for organizing nodes in a network
In current Distributed Storage Systems, applications may determine the device on which data shall be stored. The invention provides a method for automatically selecting an individual storage device within a DSS. A method for automatically classifying a node in a network comprises determining data classes, automatically classifying data of said data classes stored on a node, determining the amount of stored data per data class, calculating data dominance values for each of the data classes, receiving a request that indicates a data type according to one of said data classes, and providing said dominance value according to the requested data class. Further, a method for automatically organizing nodes in a network comprises receiving a request for storing data of a particular data type, transmitting the request to storage nodes, receiving from the storage nodes respective dominance values, and selecting the node with the highest dominance value for storing the data.
1. A method for automatic classification of a node in a network, comprising the steps of
determining at least one data class;
automatically classifying data stored on said node according to said at least one data class;
determining the amount of the data classified according to said at least one data class, wherein the determining the amount of the data classified according to said at least one data class comprises determining the number of files stored on said node which are classified according to said at least one data class and the average size of said files;
calculating, from the determined amount of the data classified into said at least one data class and a relative potential storage capacity, a dominance value for each of said at least one data class, said dominance value being a measure for the relative dominance of the data stored on the node classified according to one of said at least one data class to all other of said at least one data class;
receiving a request that indicates a data type corresponding to one of said at least one data class; and
providing, in response to the request, said dominance value for said one of said at least one data class corresponding to said data type.
2. The method according to claim 1 , wherein the amount of data is determined in terms of a combination of said file sizes and said number of files.
3. The method according to claim 1 , wherein the relative potential storage capacity of a storage device on said node is automatically estimated, said estimating comprising the steps of:
determining a data type of data already stored on the storage device;
determining file size statistics of the already stored data of said data type;
determining the remaining free storage capacity on the storage device;
receiving a request to store data of said data type, the request not indicating a file size;
according to the determined file size statistics, determining an expected size of data of said data type; and
if the remaining free storage capacity of the storage device is higher than the expected size of data of said data type, accepting storage of the data.
4. The method according to claim 1 , wherein the dominance value of a node relating to a data type is calculated according to:
E dominant,type =RPSC*A total *k 1
with
A total =N total *S total
RPSC=k 2 −k 3 ( k 4 +S average +C used )/ C total
wherein k 1 , k 2 , k 3 , k 4 are constants, N total is the total number of data files of said data type, S total is the accumulated size of files of the data type, S average is the average size of files of the data type, C used is occupied storage capacity of the node and C total is the total storage capacity of the node.
5. The method according to claim 1 , wherein the classifying and/or determining of a data class is done according to metadata associated to the data.
6. The method according to claim 1 , wherein the nodes are peers in a peer-to-peer network.
7. An apparatus for automatic classification of a node in a network, the apparatus comprising
means for determining at least one data class;
means for automatically classifying data stored on said node according to said at least one data class;
means for determining the amount of the data classified according to said at least one data class, wherein the determining the amount of the data classified according to said at least one data class comprises determining the number of files stored on said node which are classified according to said at least one data class and the average size of said files is determined;
means for calculating, from the determined amount of the data classified into said at least one data class and a relative potential storage capacity, a dominance value for each of said at least one data class, said dominance value being a measure for the relative dominance of the data stored on the node classified according to one of said at least one data class to all other of said at least one data class;
means for receiving a request that indicates a data type corresponding to one of said at least one data class; and
means for providing, in response to the request, said dominance value for said one of said at least one data class corresponding to said data type.