Bank Core Business Running Batch Storage Transformation Scheme  – Luisuantech

●Background analysis 

Batch operation requires processing a large amount of data and transactions within a specific period of time, including account balance update, interest calculation, transaction clearing and statement generation. However, at present, the processing cycle of this business is relatively long, which needs to be completed at about 6:00 of the next day. In the future, with the continuous growth of business data, it will inevitably affect the normal development of counter business. Therefore, we need to take measures to improve the efficiency of batch running business and shorten the business processing cycle. The factors that affect the business processing cycle of the storage system include the following aspects: 

●Old storage equipment 

The storage device has exceeded the normal refresh cycle of IT equipment since it was put into use. After many times of continuous upgrading, the equipment has reached the limit in terms of performance, and the failure rate has increased, and there are potential risks to reliability and stability. 

●Storage network bottlenecks 

The storage network is built with 4/8/16G Fibre Channel, while the service network is based on Ethernet technology. Data transmission between these two different network architectures must go through complex protocol conversion and data replication processes, which will undoubtedly reduce the efficiency and response speed of data transmission. 

SOLUTION

●Scheme design 

This scheme selects the LUISUAN high-performance storage solution. The storage system adopts multi-node redundancy design, and its reliability is as high as 99.9999%. This design ensures continuous, non-disruptive business operations and meets the high availability requirements of the business. 

Diskless architecture is adopted to realize the independent deployment of storage IO node and storage data node. Capacity and performance can be flexibly scaled to meet the needs of business growth over the next 3-5 years. 

High-performance NVMe SSD is adopted as storage medium to meet the requirements of high IOPS and low latency of real-time services. Ensure that data can be read or written in the shortest time, and ensure the normal operation of real-time data processing, real-time transactions and other applications. 

The storage network adopts high-performance RoCE v2 networking, which is in line with the development trend of data centers. At the same time, it also meets the requirements of easy management and maintenance to reduce labor costs and operating costs. 

●Programme value 

●High availability 

Full redundancy design and automatic switching mechanism of failed links can minimize the risk of business interruption caused by storage system failure. 

●Business acceleration 

Provide ultra-high-speed data sharing resource pool to meet the rapid access and access of different business data. With extremely low latency, the data transmission time of the batch service is reduced to the extreme, thereby effectively improving the service data transmission speed. 

●I/O boost 

It provides tens of millions of IOPS, with extremely high random read and write performance, ensuring fast data read and write, effectively improving the response efficiency of batch service data I/O read and write requests, and improving service response efficiency. 

RELATED PRODUCTS

LUISUAN LinePillar FS Parallel File System 

●Safe and efficient 

Metadata nodes support Active-Active pairing mode, which realizes the mutual backup of metadata among metadata nodes to ensure the continuity of metadata service. Support global erasure code, storage space utilization rate can reach more than 90%. 

●Large amount of small file optimization 

Support unified storage and efficient retrieval of tens of billions of files. Innovative small file Container storage technology improves the efficiency of small file creation and retrieval to more than ten times that of traditional storage. 

●Rich access interfaces 

Provides a comprehensive set of storage interface protocols including file storage (POSIX, NFS, CIFS), block storage (iSCSI), object storage (S3, Swift), big data (HDFS), and container storage interface (CSI). 

ForinnBase GroundPool 5000

EBOF  ForinnBase GroundPool 5000 EBOF (GP 5000), which enables offloading and encapsulation of data protocols by perfectly integrating current high-speed flash transfer protocols with storage-specific ASIC chips. It has the characteristics of low delay, low power consumption, high throughput, large capacity and easy expansion.