laitimes

Case | Bank of Communications Successfully Implemented Data Redundancy Solution, Accelerating Wide-area Expressway

author:Digitization of finance

Text / Bank of Communications Data Center, Dai Kaiming, Liu Peng, Yao Li

Science and technology are the primary productive forces, and scientific and technological innovation is the strategic support for improving social productive forces and comprehensive national strength. At present, a new round of scientific and technological revolution and industrial revolution is developing in depth, new technologies are bringing subversive changes to the economy and society, and the leading and supporting role of the digital economy in economic and social development is becoming increasingly prominent. As the most intensive field of technology application, the financial industry actively grasps the opportunities of the development of the digital economy, and continuously promotes the high-quality development of the industry through technological innovation and cutting-edge technologies such as 5G, cloud computing, and artificial intelligence.

Behind the digital transformation of the financial industry is the high-speed interaction of massive data between multiple data centers and branches, and the financial WAN is the core artery to realize this interaction. In order to consolidate the foundation of digital technology and improve the continuity of IT services, Bank of Communications has started the construction of two data centers in Inner Mongolia, Lingle New Area and Gui'an New Area in Guizhou, and the computing power system is evolving from "three centers in two places" to "multiple centers in multiple places". At the same time, based on IPv6+ technology, we will build core backbone networks, build a new generation of hierarchical, modular, and intelligent enterprise-level financial cloud WANs, and build an information superhighway for digital transformation.

Three factors drive the continuous growth of WAN traffic

With the deepening of digital transformation and the continuous expansion of inclusive finance and audio and video services, the data transmission of financial cloud WAN is driven by three major factors: first, business growth drives the continuous growth of wide-area traffic; second, the transformation from centralized to distributed architecture, and the demand for batch data replication and transmission between multiple data centers surges, resulting in a contradiction between a large number of data transmission requirements and limited private line bandwidth resources; and third, the diversification of financial services, and the transformation of bank outlets from "single service" Diversification of interaction modes such as online and offline integration, as well as transformation of ecological services such as government affairs and people's livelihood. These factors lead to increased pressure on WAN transmission, the need for continuous expansion, and the need for greater cost pressure.

Case | Bank of Communications Successfully Implemented Data Redundancy Solution, Accelerating Wide-area Expressway

Figure 1 Distributed cloud construction drives the rapid growth of WAN traffic

In 2021, BOCOM and Huawei have been working together to explore a next-generation WAN redundancy reduction solution to reduce the size of service data in WAN transmission. In various business interaction scenarios such as between data centers, branches and head offices, Bank of Communications conducted compression tests for different business types and time periods, fully verified the compression effect in different scenarios, and continuously optimized the average compression ratio and the high-availability architecture of the compression scheme. In October 2023, Bank of Communications' next-generation enterprise-level financial cloud WAN was deployed in the Hefei-Shanghai section on a priority basis, with stable production and operation, with an average traffic compression ratio of 2.9:1 from Hefei Center to Shanghai Center, a bandwidth transmission efficiency improvement of about 65%, a long-distance bandwidth demand reduction of about two-thirds, and a transmission delay increase of less than 0.1 milliseconds, without any intrusive transformation requirements for application systems.

Case | Bank of Communications Successfully Implemented Data Redundancy Solution, Accelerating Wide-area Expressway

Figure 2 Bank of Communications' next-generation wide-area data redundancy solution

Five major innovations enable efficient transmission of financial data

The origin of data compression can be traced back to Shannon, the father of information theory, who proposed Shannon coding in 1947, which plays an important role in storage systems and modern data communication. From simple file compression to complex audio and video data processing, data compression technology is relied on to reduce the amount of data required for information storage and transmission, thereby saving storage space and speeding up data transmission efficiency. With the development of technology, new compression algorithms continue to emerge, which can process large amounts of data more efficiently and improve compression efficiency. By eliminating data information redundancy, the packets are compressed and transmitted after being compressed, so that the actual number of packets transmitted in the same time is increased compared with the number of packets before compression, so as to improve the transmission efficiency of the WAN. Compared with the traditional point-to-point data compression solution, the wide-area data redundancy elimination solution of Bank of Communications has five major innovations:

1. Efficient compression to meet the growth of backbone network traffic: Based on the mainstream ZSTD equalization algorithm and stateless protocol innovation, WAN data redundancy reduction reduces the data transmission bandwidth between DCIs by 35%~70%, saving a lot of private line bandwidth costs every year. The additional delay generated by the compression and decompression process is controlled at the order of 0.1 milliseconds, which is negligible compared with the total delay of tens of milliseconds in the whole transmission process of the backbone network, and minimizes the perception of upper-layer services.

2. Board compression to meet simplified deployment and high availability: WAN data redundancy is based on a built-in dedicated data redundancy board in the egress router, eliminating the need for additional compression equipment, and completing data diversion, compression, and forwarding in one stop, with simple networking and protocol agnostic. At the same time, it supports inter-board redundancy and inter-device backup, eliminating data redundancy and eliminating business impact on board failures, meeting the requirements of high financial availability and continuity.

3. Lossless compression to meet data consistency requirements: Traditional compression protocols lose part of the five-tuple information during the compression process, which loses the refined management of business flows. WAN data redundancy retains the original packet header information during the compression process, which does not affect the refined identification and statistics of service flows and the execution of QoS policies during data forwarding, and does not damage any service data.

4. Multi-point compression to meet large-scale flexible networking: The mainstream data compression scheme for point-to-point private lines is not suitable for large-scale networking scenarios from multipoint to multipoint. After data compression, it can be sent to any destination according to business requirements, and the destination device identifies the compressed package and decompresses it to meet the requirements of distributed multi-point interconnection scenarios.

5. Network-level compression to meet the needs of non-perceptual applications: The WAN data redundancy elimination solution implements IP packet compression between network devices, which is transparent and imperceptible to application interaction, does not require any upgrades or transformations on the application side and database, and is easy to quickly deploy and go online.

In the face of increasing pressure on business traffic, BOCOM will continue to innovate by exploring better dictionary compression algorithms to further improve the data compression effect, and on the other hand, it will gradually promote data redundancy reduction solutions between data centers and branches to data centers to continuously reduce costs and increase efficiency. At the same time, BOCOM will also explore the integration and application of cutting-edge technologies such as iFIT, 5G cloud private network, digital map, and quantum security with financial cloud WAN, continue to optimize the basic network architecture, improve digital operation methods, and build a solid foundation for "building a world-class banking group with distinctive advantages".

Read on