Abstract:Huawei Cloud GAUSSDB (for OpenGauss), a new generation of financial distributed database, has officially launched several major new kernel features, such as USTore storage engine, DCF high availability component based on Paxos protocol.

In the digital era, the technological iteration and update are more rapid than ever before, and the database technology also spurts out a huge development tension in this technological wave, and the core technology, as the soul of the database technology, also realizes the leap type development.

Huawei cloud database has always attached great importance to kernel technology research and development, has a wealth of database kernel research and development experience, and continues to build a series of highly competitive kernel features. Recently, Huawei Cloud GAUSSDB (for OpenGauss), a new generation of financial distributed database, has officially launched USTore storage engine, Paxos-based DCF high-availability component and other major new kernel features, creating a stable and high-performance storage engine and high-availability kernel capability for enterprise users. Taking innovative technologies as the starting point, we will further promote the digital transformation of enterprises.

USTore storage engine innovation, frequent data updates under the performance is still as stable as Mount Tai

USTORE storage engine, also known as In-place Update storage engine. In-place Update is a new storage mode In GaussDB(for OpenGauss) kernel. The GAUSSDB (for OpenGauss) kernel currently uses the row engine APPEND UPDATE mode. Append updates work well for business additions, deletes, and HOT(Heap only tuple) updates (updates within the same page), but garbage collection is not efficient for non-HOT UPDATE scenarios across data pages.

And Ustore engine will be the latest version of the “valid data” separation and historical versions of junk data storage, the latest version of the “valid data” is stored in the data page, and open up a UNDO space alone, used for unified management history version of “garbage data”, so the data space is not due to the frequent updates and expansion, garbage collection efficiency is higher. USTore engine adopts the Undo subsystem design of NUMA-Aware, which enables the Undo subsystem to be effectively extended on multi-core platform. At the same time, multi-version index technology is adopted to solve the problem of index cleaning and effectively improve the efficiency of storage space recycling and reuse.

Figure 1: USTORE data storage schematic: Data is on the original page, and Undo is managed separately through a pointer

Overall, the USTORE data store delivers higher performance, higher efficiency, better space utilization, and smoother overall system performance to accommodate a wider range of business scenarios and workloads than the original garbage data cleanup mechanism.

In terms of practical application, USTORE storage engine has been POC tested in some customer sites, showing stable and reliable database service ability, and achieved the expected results.

Innovative high-availability DCF components based on the Paxos protocol ensure strong consistency and high availability of data

In addition to the USTore storage engine, GAUSSDB (for OpenGauss) has also introduced a Paxos-based DCF high availability component, which enables GAUSSDB (for OpenGauss) to further enhance its high availability while ensuring data consistency. These components include:

(1) Get rid of the third party arbitration component through self-arbitration and majority selection, greatly shorten the RTO time, and can prevent the split brain under any fault;

(2) Multi-cluster deployment mode which supports synchronous and asynchronous mixed deployment of nodes;

(3) Improve the log replication efficiency between primary and standby nodes, and improve the maximum throughput capacity of the system.

Figure 2: Schematic diagram of Paxos

With the DCF high availability component of GausSDB (for OpenGauss), users can not only avoid the risk of a system brain-splitting, but also improve availability.

In the future, GAUSSDB (For OpenGauss) will continue to build more intelligent, safer and more efficient technical capabilities based on AI autonomous technology in USTore storage engine, Paxos protocol optional master and other key features, and proceed from the most pressing problems and long-term needs of database applications in thousands of industries. Continuously in the core technology breakthrough, for customers to create a more leading, more high-quality database services.

[Highlight of Events]

On August 18th, the 12th China Database Technology Conference (DTCC2021) with the theme of “Digital Building the Future” will be held in Beijing. A special session of Huawei Cloud GAUSSDB will be held on the scene. GAUSSDB series database and core capabilities will be explained for you, and the best practices of GAUSSDB industry will be shared. Product UGO GaussDB, moreover, the migration tool is currently in beta, welcome to huawei cloud website free experience > > https://www.huaweicloud.com/p…

Click on the attention, the first time to understand Huawei cloud fresh technology ~