Disaster tolerance system based on block chain technology, can provide the disaster data center with higher availability, can also according to the need to build an industry-wide disaster chain alliance, through mutual cooperation to build strong resistance against natural disasters, it is one of the main development direction of distributed storage,
FileStorm as a support for disaster preparedness storage distributed storage platform, disaster preparedness nodes all over the world, it provides a connection enterprise database middleware FileStormDR production node and disaster preparedness, system by extracting the local database log operation, reached consensus on the node and stored in the form of trading on FileStorm, more can not only realize the database node backup, may be conducted through the log operation restore the database to any point in time,
FileStormDR production node as a connection enterprise database and disaster preparedness storage nodes in the middle of the plug-in, not only suitable for FileStorm, can also be applied to a variety of distributed storage based on block chain implementation platform,
Technical principle
1. In the center of the production, the client application to the database, operation records recorded in the database log file (Binlog, RedoLog, etc.),
2. The middleware production node through the analysis of the API interface to read FileStormDR production database operations of data in database log file encapsulation into suitable for block chain network transaction data,
3. Production node will forward transaction data to other middleware backup node, and temporary storage down ready to write blocks of middleware,
4. Middleware between all nodes consensus for temporary storage of transaction data in each node is formed on the same block, the block contains the same content in the same order, after dealing with the consensus of each node on the trading data agreed,
5. Backup nodes participate in block chain network consensus, have the same data, and other nodes when generate new blocks, the backup node read records in the block trading data and sent to disaster preparedness center database,
Figure 1
Technical features
FileStorm as can support a variety of mainstream database backup, such as, essentially, MySQL, Oracle, DB2, MangoDB, HBase, SQLite,
Can be produced in 10 seconds will center data is passed to the disaster preparedness, and through the log replay to disaster database, make disaster preparedness center RPO (Recovery Point Objective) control level in 10 seconds,
Support production center and disaster preparedness center one-to-one and one-to-many mode of deployment, many disaster preparedness center to back up your data at the same time, due to use the underlying technology of block chain, to ensure that each disaster data at the same time be confirmed and write in the center of the middleware, maintain strong data consistency between the multiple copy, copy number will have little impact on the RPO, basic negligible,
Advantage
Expand accessibility to node
FileStorm can convenient online disaster added node at any moment, does not affect the original disaster preparedness system, add the disaster preparedness nodes automatically obtain the original log information from other nodes, and the local playback log, generate new disaster database,
Audit, operation records do not tamper with the
FileStorm underlying technology based on block chain, perfect inherited block chain technology tamper-resistant features, all the database DDL and DML operations are detailed records in the middleware, do not change, and delete data can be accurate tracking, query, provides a complete database audit basis,
Online upgrade and maintenance
In disaster preparedness center more cases, FileStorm can take online polling way to upgrade the software and hardware of each disaster preparedness center, the upgrade process will not on production centres and other disaster preparedness center business have any effect,
Hardware specifications without consistency requirements for disaster preparedness center
Disaster preparedness center can according to the actual need to configure the corresponding hardware, this characteristic is very suitable for distributed storage based on block chain implementation platform, hardware specifications inconsistent on the middleware disaster efficiency impact almost negligible, the user can also use both old equipment for deployment in the center of the disaster, integrate existing resources, make full use of various resources to avoid the waste of resources,
Perfect security mechanism
In order to ensure production center and disaster preparedness center writing data and the security of data transmission, FileStorm provide symmetric and asymmetric encryption algorithm, users are free to choose,
Perfect management tool
FileStorm will provide a graphical interface management and monitoring tools, data backup and recovery become simple and easy to use, FileStormDR can develop into a separate storage management system, can be applied to centralized storage and the second backup storage is going mainstream,
In this paper, all of the data, from the test environment as follows:
The operating system
Linux (CentOS, RedHat enterprise edition, Ubuntu 15.04 +, SUSE) Windows 7 +
Hardware
CPU: 64 - bit x86_64, 4 + cores
Disk: 200 gb SSD
16 GB + RAM:
Author: block chain _ offer brother