Performance leads the cloud native database market! Intel and Tencent jointly build cloud technology ecology

  In recent days, , Tencent cloud native database TDSQL-C Usher in a comprehensive upgrade . The new version covers Intel based   Haughty   L2 cache of persistent memory 、 Full link RDMA The Internet 、7*24 Core functions such as continuous lockless backup , In the full cache scenario and the big data set scenario , Performance improvement compared with traditional cloud database 200%, Beyond other cloud native database products in the industry . meanwhile ,TDSQL-C The cost performance has also been greatly improved , The price is only... Of the commercial database 1/10

                                                        Optimize product architecture in multiple directions

                                                        Intel and Tencent jointly innovate

  As a new generation of cloud native relational database developed by Tencent cloud , This release of TDSQL-C The new version is equipped with Tencent self-developed database kernel TXSQL,TXSQL A lot of optimizations have been made for cloud native database scenarios , It can greatly improve the database performance , Better meet the enterprise's pursuit of database performance .

To further optimize the product architecture and technical implementation ,TDSQL-C Actively carry out joint innovation with Intel technology team . Combined with the latest generation of Intel   To the strong   Scalable processors and Intel   Haughty   Persistent memory (PMem) The design scheme of reconstructing L2 cache based on the hardware characteristics of , New version in IO bound The read / write performance in the scenario is improved 2 More than times .TDSQL-C We also work with Intel to optimize the design of storage solutions in multiple directions , For example, join polling 、 Algorithm optimization 、 Eliminate locking and other mechanisms , Optimize the storage engine, etc , And introduce the... Provided by Intel SPDK Development Kit , Optimize NVMe Solid state disk IOPS And delay performance .

Besides ,TDSQL-C The new version also optimizes and upgrades the network architecture . Intel   Haughty   Full link of persistent memory RDMA The Internet , Rely on zero copy 、 Kernel bypass 、 nothing CPU Intervention, etc , Storage layer 、 The system performance of the critical path between multiple copies of the computing layer and the storage layer is further optimized , Reduce request latency by up to 80%, send I/O Performance is no longer a bottleneck .

                                                        Excellent performance in multi scenario tests

                                                            Obvious performance advantages

  In exploring the actual performance of the new version , The tester passes the benchmark tool sysbench And full simulation business production environment , Write only for each 、 The read-only and mixed read-write scenarios are tested .

According to the report TDSQL-C Leading similar products in the market in terms of performance , It can fully meet the requirements of enterprises for database performance . The test data in the report is only the result of a single node ,TDSQL-C MySQL Support second level extension , It can help enterprises improve their reading performance to 1500% above .

in addition , In the hot data access scenarios of e-commerce business and game business , It is usually necessary to put all the data into the cache , And the disk update cache does not need to be read or written during the query process . The test report shows , In the full cache scenario , The larger the instance size TDSQL-C MySQL The more obvious the performance advantage of , stay 32 After nuclear specification ,TDSQL-C MySQL Version can be added CPU To further improve QPS. At the level of resource utilization ,TDSQL-C MySQL Version can calculate the CPU Utilization reaches 90% above , At request delay rtt There is almost no shaking on the .

Compared to the full cache scenario , Big data set scenarios are more frequently used in real business . In the read-only scenario of large data sets , In the general scenario of mixed reading and writing TDSQL-C MySQL The performance of this version exceeds that of similar products in the market , And as the size increases , The performance advantage is more and more obvious .

Cloud native database TDSQL-C MySQL Version adopted “ Storage and computing are separated ” Design concept of , Support maximum super 400TB Single database mass data storage . With the application of this feature , More and more enterprises no longer use the method of dividing databases and tables to deal with the storage of historical archived data , So there is “ Super large watch ” The phenomenon of . In the single table 40 Billion records and data storage space reaches 1TB In the scene of ,TDSQL-C MySQL The version is still outstanding .

In a mixed read-write scenario , As the database size increases ,TDSQL-C MySQL The performance of version is better than that of traditional cloud database MySQL Promoted 200%, It also has obvious advantages over other cloud native databases .

 

                                                              Push the technology to the ground

                                                          Build cloud primary ecological integration

At present, with the blessing of Intel Technology ,TDSQL-C It has provided stable and reliable enterprise level cloud database services for users in different industries , With backup 、 recovery 、 monitor 、 The rapid increase 、 Data transmission and other database operation and maintenance of a complete set of solutions , Can help enterprises simplify IT Operational work , Focus more on business development .

future , Tencent cloud database kernel team will continue to carry out in-depth cooperation with Intel and the open source community , Continue to contribute technical strength , Bring new business value and technical advantages to all walks of life , Spare no effort to promote co construction of production and research and ecological integration .