Shulou(Shulou.com)05/31 Report--

This article is about how to improve distributed performance in Oracle12c. The editor thinks it is very practical, so I hope you can get something after reading this article. Let's take a look at it with the editor.

The recently released version of Oracle 12c R2 supports database sharding, which is intended to improve the performance of Oracle's flagship database in a cloud architecture environment and benefit from a database scale-out framework.

Improving data scalability in a cloud environment is important for Oracle because Oracle takes on more load and requires lower costs as it migrates from an internal build environment to a cloud environment.

Oracle sharding draws on the strengths of NoSQL databases such as Cassandra, MongoDB, and Amazon DynamoDB. These and other NoSQL databases support sharding technology, allowing databases to be deployed horizontally across multiple servers.

This technology is not only limited to NoSQL, but also supported in many mainstream relational databases, such as PostgreSQL and MySQL databases.

At the Oracle database route map seminar at Oracle 2016, Andrew Mendelsohn, executive vice president of database server technology at Oracle, said: "Amazon has been using sharding technology for many years." However, he accused the implementation of Amazon of requiring too much human intervention.

There are many sharding schemes that require application code to control a lot of activity. Mendlesohn said that Oracle uses the local database sharding architecture, which reduces the control work of the database management system and saves development time.

He added that Oracle's initial strategy was to support sharding across a thousand servers, but that support was expected to expand.

Server farm environment slicing application

Slicing technology is widely used in the data field of Facebook running open source relational database MySQL database. This may be surprising, but the social media giant also uses Oracle databases for financial and operational applications. Sudhi Vijayakumar, a database systems architect at Facebook, mentioned this information when he communicated at the sharding technology section of the Oracle Global Conference.

Vijayakumar has worked with Oracle engineers and has been preparing the proof-of-concept implementation of the data Mart, which will achieve quasi-real-time report generation. "We look forward to Oracle implementing sharding technology," he said.

He suggests that Oracle shredding can support the growing needs of applications without the need for additional staff to manage it. Usually, traditional enterprises pay more attention to this goal, but it is rare for benchmark enterprises such as Facebook to pay attention to this aspect.

Oracle innovation is in a dilemma.

Oracle sharding can use the company's RAC (real-time application cluster) or not. These clusters use shared disks to provide familiar relational data consistency, and Oracle has extended databases in this way for many years.

Many people think that the association cost of the RAC approach is too expensive, and in many cases, software architects build very large-scale clouds. Moreover, most of the work can be satisfied through "ultimate consistency". In this way, companies can make extensive use of sharding technology while creating many distributed data technologies.

Guy Harrison is an independent consultant and writer based in Melbourne, Australia. He believes that sharding technology is a better architecture than RAC for large-scale expansion. RAC relies more on scale-up than scale-out; RAC requires a shared disk system independent of database nodes.

"with sharding, each node has its own disk, which may be directly on a commercial server," he said. "if you need to add nodes, you can add disks, CPU and memory, and use cheaper commercial 'pizza box' servers." (pizza box server here means ordinary cheap server.)

He says there is less collaboration between nodes, and of course it depends on the application. The disadvantage of this architecture is that the query efficiency within the cluster is low, and the consistency is slightly poor, so it is difficult to maintain true ACID consistency across clusters.

Harrison has little doubt about the overall trend of migrating applications and databases from on-premises to the cloud. He said that the database is the part that is migrated to the cloud, and the database is usually considered according to the situation of the application.

"in many cases, when people want to build something new, they will make a comparison between spending a lot of money internally buying a lot of hardware and getting convenient services from the cloud," he said. it's not a very difficult decision. compared to the environment you can rent in the cloud, building these infrastructure is very expensive. "

Harrison says he is interested in migrating Oracle databases to the cloud. However, like others, he also expects the corresponding economic benefits of moving to the cloud.

"to some extent, they want to get better revenue than on-premises deployment. without any benefits, it is difficult to get their users to migrate to the cloud. So, this is a thorny problem."

The above is how to improve distributed performance in Oracle12c. The editor believes that there are some knowledge points that we may see or use in our daily work. I hope you can learn more from this article. For more details, please follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.