The following article mainly analyzes the current situation of MySQL databases with internal and external concerns. We all know that Oracle's acquisition of Sun in order to obtain MySQL is a painstaking effort, the final approval from regulatory authorities. Currently, these efforts may be a waste of time and money.
People may suddenly find that MySQL is on the verge of extinction.
At the MySQL conference held last month, Michael wenius, the father of MySQL, and BrianAker, the famous MySQL architect, delivered speeches respectively, they firmly believe that no company can become the only provider of MySQL development or support services. The practice of these MySQL database celebrities is a test for Oracle, which will verify the degree to which Oracle and MySQL Community cooperate and tolerate different opinions.
Clustrix, a new company in San Francisco, recently announced that its products are better and better, and it can complete MySQL's poor operations. It can be expanded to store billions of pieces of data and replace MySQL.
The Clustrix product does not contain MySQL DNA, but it can communicate with the MySQL Protocol, so that the application no longer needs to be transplanted code, and its existence will undoubtedly harm the paid business of the MySQL database.
This product is called the first cluster Database System for Internet-scale applications. It is said that it complies with the evolution of application servers and storage systems to scalable and clustered products.
It has the huge scalability and high performance of NoSQL key/value storage, and the SQL encapsulated in the 3-node server CLX 4010 device has reliable ACID test functions, this hardware device is sufficient to handle high-load read/write data operations.
All three or more rack-mounted devices need to run a software called the siider cluster database engine. According to Clustrix, how much extensibility a user wants or needs depends on how many node devices are added to the rack.
The siider cluster database engine is a non-shared execution environment, including the ParallelPlanner and Distributed ExecutionEngine ). It provides the query task to distributed data, rather than the data provided to the query task like RDBMS.
This means that the Clustrix cluster database should be able to execute query statements in the maximum concurrency, and many synchronous queries have the maximum concurrency. This will bring about extremely high scalability, read/write operation performance, availability, online adjustment outline, self-repair and self-management.
The Clustrix team has learned a lot from Isilon Systems, which has developed similar high-concurrency and distributed products for storage Systems. Clustrix has obtained $18 million from venture capital institutions to develop scalable databases.
The target user group of the Clustrix cluster database is a cloud computing service provider for transaction processing, enterprises, and Internet companies such as social networking websites. When they process incredibly tedious data in Internet life, to obtain the needed scalability, you have to endure the database merging and database splitting operations at the application layer. The open-source projects Hadoop and Cassandra, as well as Google's BigTable, are also solving the problem.
To expand the MySQL database, it usually requires a lot of painful customization programming. This is a high cost and time-consuming task, and it is difficult to find the relational database function on the Internet in a single instance database. Clustrix promises that with its products, people no longer need to write such code.
The Clustrix cluster database system can be scaled up to hundreds of nodes in an incremental and seamless manner. It runs just like a single instance database, with all relational database functions and consistent real-time transaction processing.
This tool can be deployed to the sharding, non-sharding, and replication MySQL database environments transparently and without interruption. When the customer needs to add more CLX4010 nodes, the distributed and parallel architecture can automatically release data to new nodes, even if the data writing load is very heavy, it can also linearly improve performance. It achieves high availability through automatic load balancing, failover, restoration, and self-repair.
Clustrix dual-core and quad-core devices include two 1 Gbit/s Ethernet ports and two 20 Gbit/s InfiniBand backplane ports. They are also equipped with 32 gb ram and 7 gb ssd. The price for a three-node device is USD 109995.
Clustrix said it had previously started selling these products and reached its first deal in the first quarter of this year. It is said that the company has received a lot of orders, which is obviously attractive to some MySQL customers.