Rapid fault node repairing method based on bandwidth perception
A technology for faulty nodes and repair methods, applied in digital transmission systems, electrical components, transmission systems, etc., can solve problems such as service quality impact, time consumption transmission time, recovery time impact, etc., to improve deployment and repair efficiency and improve availability. and reliability effects
Active Publication Date: 2021-04-27
CENT SOUTH UNIV
View PDF11 Cites 2 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
In large-scale storage systems, storage node failures occur very frequently, and the long recovery time may affect the service quality or even affect the entire system.
Due to differences in actual bandwidth between servers, most of the time spent on completing node repair is spent on data transmission time
The traditional repair process randomly selects servers, which further prolongs the repair time and leads to waste of resources
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0035] Such as figure 1 It is a schematic flow diagram of the method of the present invention. This bandwidth-aware based faulty node fast repair method provided by the present invention comprises the following steps
[0036] S1. Obtain the original data, divide it into k data blocks, and specify the size of each data block as ω; when encoding k pieces of data, generate n-k pieces of encoded data, a total of n pieces of data. The n copies of data are respectively stored in n storage nodes of the erasure code storage system. If any piece of data (or a node) fails, you only need to randomly select k pieces of data from the remaining n-1 pieces of data to perform decoding operations to restore the failed data.
[0037] S2. Design the generation matrix according to (n, k) erasure codes, code k original data blocks to generate n-k check data blocks of the same size, and stipulate that the size of each data block is ω; a set of n data blocks is called a Band, distributed in the...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a rapid fault node repairing method based on bandwidth perception. The method comprises the steps that coding operation is conducted by obtaining original data; a source point server is set, and a data packet is sent to an adjacent server by the source point server to confirm the real-time bandwidth of a link between the servers; the source point server constructs and deploys a data transmission link according to the real-time bandwidth condition fed back by the communication server; deployment pre-detection is carried out, and an erasure code repair network is constructed for the failure nodes; and the server storing the failure data block is set as a failure server, repair pre-detection is performed according to the real-time bandwidth condition fed back by the source point server, and the erasure code repair task distribution condition is optimized. According to the method, the server with the optimal bandwidth is dynamically selected to participate in the task of deploying and repairing the erasure codes, so that the load balance of the whole storage system can be ensured, the efficiency of deploying and repairing the erasure codes is effectively improved, and the usability and reliability of the system are further improved.
Description
technical field [0001] The invention specifically relates to a fast repair method for faulty nodes based on bandwidth perception. Background technique [0002] In recent years, the amount of data stored in large-scale distributed storage systems has grown rapidly. In traditional storage systems, triple copy technology is a commonly used reliability mechanism, which uses copies to replace failed nodes to complete fault recovery. This method is simple and easy to use, but it introduces unaffordable storage cost and triple storage overhead, which makes reducing storage overhead an inevitable task in large-scale storage systems. Many storage systems have started to use erasure coding as their reliability mechanism. As a widely used erasure code, RS (Reed-Solomon) code works as follows: when data is written into the storage system, the original data will be divided into k data blocks of the same size to jointly generate a matrix to complete the encoding process . The generati...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): H04L12/24
CPCH04L41/0654
Inventor 朱兵赵旭煜曾志伟王伟平王建新
Owner CENT SOUTH UNIV