Storage network structure based on the Peterson graph and data read-write method thereof

• US 20100077101A1
• Filed: 09/11/2009
• Published: 03/25/2010
• Est. Priority Date: 09/12/2008
• Status: Active Grant

• Current Assignee: Beijing Acell Technology Company Limited, Institute Of Acoustics Chinese Academy Of Sciences (Chinese Academy of Sciences)
• Original Assignee: Beijing Acell Technology Company Limited, Institute Of Acoustics Chinese Academy Of Sciences (Chinese Academy of Sciences)
• Inventors: Wang, Jinlin, Wang, Lingfang, Li, Tingyi, Qi, Xiangdong, You, Jiali
• Granted Patent: US 8,219,707 B2
• Time in Patent Office: Days
• US Class Current: 709/238
• CPC Class Codes: H04L 67/104 Peer-to-peer [P2P] networks
• CPC Class Codes: H04L 67/1065 Discovery involving distributed pre-established resource-based relationships among peers, e.g. based on distributed hash tables [DHT] pre-configuration of logical or physical connections H04L67/1053
• CPC Class Codes: H04L 67/1076 Resource dissemination mechanisms or network resource keeping policies for optimal resource availability in the overlay network

Abstract

The present invention provides a storage network structure based on Peterson graph and data read-write method thereof. The aforementioned storage network structure consists of a primary distributed storage network and a Peterson graph-based storage network, in which, the said primary distributed storage network may be a cluster distributed storage network or a DHT (Distributed Hash Table)-based peer-to-peer storage network consisting of n nodes; and the said Peterson graph-based storage network is formed by 10 nodes selected from aforementioned storage node set according to policies. The Peterson graph-based storage network structure and data read-write method thereof may provide the different reliability classes for the applications by combining peer-to-peer based distributed network storage with Peterson graph-based network storage, solve the problem of a large quantity of data migrating by utilizing the better media data migrating non-movable property from the Peterson graph-based network storage, and screen the single failure of the cluster storage structure and avoid performance degradation of peer-to-peer wide area storage by combining with the availability and robustness of Distributed Hash Table.

Claims

1. A storage network structure based on the Peterson graph, is characterized of including a primary distributed storage network and a Peterson graph-based storage network,

wherein,

the said primary distributed storage network is the cluster distributed storage network or the peer-to-peer storage network based on the distributed hash table, consisting of n storage nodes, wherein n is the integer more than 10;

and the said Peterson graph-based storage network is formed by 10 nodes selected from the set of the said storage nodes according to the policies, in which the degree of each node is more than or equal to 3 and the distance (hop) between any two nodes is less than or equal to 2, where the said policies are the storage capacity, the available bandwidth, the processing capacity, and the degree of the node or the distance between nodes.

2. A data read-write method of a storage network structure based on the Peterson graph, in which the said storage network structure based on the Peterson graph includes a primary distributed storage network and a storage network based on the Peterson graph, is characterized of including a method used by applications to write data and a method used by applications to read data, in which the said applications are the entities except the storage network, such as a software, an application program or a client program,

and the said method used by the applications to write data includes following steps:

1) the said applications send a data-write request carrying with the QoS (quality of service) parameters, wherein the said request includes the QoS parameters, the data identity and the data, the said QoS parameters are defined as two metrics, the data reliability metric and the data type metric, each of which may be defined into the several classes;

2) the said storage network receives the said data write request, which one central node of the storage network or one node of the distributed storage network receives the said data write request sent by the said applications,

wherein the said central node refers to a specific server addressed by dns (domain name service) or the alternative addressing method, and all of the data write requests are delivered to this central node;

3) the said storage network resolves the said data write request, which includes resolving the QoS parameters into the data reliability metric R_application and the data type metric T_application; and

4) the said storage network executes data-writing, which includes the operations of metadata writing and media data writing by the said primary distributed storage network, wherein in case of R_DHT<

R_application, the Peterson graph-based storage network executes both metadata writing and media data writing, and the said R_DHT is the reliability metric of the primary distributed storage network,

wherein the said data read method consists of;

1) the step of addressing and locating the metadata, which may be performed simultaneity in the primary distributed storage network and the Peterson graph-based storage network, wherein the first returned response is the result; and

2) the step of reading the media data, in which after obtaining the metadata, in the case that the media data contains the multiple storage nodes, the performances in terms of the delay and available bandwidth among these storage nodes are compared, and the optimal node is chosen to fulfill media data reading.

3. The data read-write method of the storage network structure based on the Peterson graph according to the claim 2 is characterized in that

in the said step 1) of the said application data write method, the data reliability metric and the data type metric of the said QoS parameter are defined as the most significant 16 bits and subsequent 16 bits of a 32-bit word, respectively.

4. The data read-write methods of the storage network structure based on the Peterson graph according to the claim 2 is characterized in that,

in the said step 4) of the said application data write method,

in the case that the Peterson graph-based storage network executes metadata writing and media data writing so as to store the data, it needs to determine whether the node ids of the storage nodes which are responsible for data that will be stored are equal between the primary distributed storage network and the Peterson graph-based storage network, wherein if equal, one of 6 nodes which are two hops away from the current node is selected from the Peterson graph to store data;

otherwise, data may be directly stored in the node which is computed from the Peterson graph-based storage network; and

in the case that T_application is the type of minimal data migration, one node in Peterson graph based storage network is selected according to measurement algorithm.