Improve quality with intelligence financial imaging platform solution

Finance

Image platform solutions

01 background

Background Overview

Commercial finance is in the process of daily business informatization , A large number of reports need to be processed almost every day 、 bill 、 Electronic documents 、 picture 、 Audio 、 Unstructured data such as video . so to speak , The management of unstructured data is an extremely important part in the process of commercial and financial informatization . Keep up with the challenging and rapidly changing times , We need to realize modernization . To keep up with and exceed changing customer needs while reducing costs 、 Changing regulatory requirements , And cope with the increasing network security risks , A huge transformation is needed .

Based on this , Commercial and financial users may develop by themselves , Or buy unstructured file management systems and financial tools , To realize the centralized management of unstructured data , Provide unified query for all financial business systems 、 Review 、 Modification and other service functions . The application of data and value mining show a huge mechanism in financial transformation , And concentration 、 convenient 、 An easy to manage unified storage platform can help financial users better realize data , Especially the regenerative application and value mining of unstructured data .

Demand analysis

Rich interfaces , Meet the historical data input in different ways The business model of commercial finance is based on outlets , Data sources are scattered , Include ： Scan input of historical data ; Desk top acquisition device input and recording of counter business 、 Video input ; Scanning of self-service devices at outlets 、 Photo data ; Outlet to superior outlet 、 Upload data from the headquarters . Each mode has its own interface mode , The storage system should provide richer interfaces to ensure that it can receive data input from different sources .

It is difficult for all business systems to exchange data The business systems of commercial finance history construction are mostly based on relational databases , Store their own business system data independently , It's hard to communicate , Therefore, the allocation of resource occupation is uneven , And there are too many data repeatedly archived , Cause a great waste of resources .

Most of the existing image systems use independent storage , Waste of resources In the traditional construction mode, each system has an independent storage platform , Space cannot be shared , Data cannot be exchanged , The value of historical documents caused by this cannot be mined and used again , A great waste of resources .

The existing mode is inefficient , Restrict the further development of business The equipment cannot cooperate with each other , Whether to use 、 management 、 Maintenance needs to be completed independently , It not only increases the burden of daily operation and maintenance , It limits the access of businesses to stored data , Business collaboration cannot be achieved , Thus, it restricts the possibility of further development among financial businesses .

Lay a storage foundation for the construction of financial big data The application of big data has increasingly promoted the financial industry , Financial users of all sizes are trying big data business .XDFS Distributed storage cluster is just for big data 、 Clouds and other emerging IT Architecture comes from , At the same time, it can meet the needs of traditional businesses , It has laid a solid storage foundation for users' future big data construction .

02 The project design

Program Overview

With XDFS Distributed file systems act as shared storage pools , It can effectively solve the problems of data storage and management ： Form a large file system with multi node clusters , Take unified namespace as the exit , Provide basic storage pool support for image management platform , Provide shared storage support for the front-end image business system . Through the high-speed network as a link , Achieve data balance within the cluster and support image storage , Concurrent response to read and write requests for front-end business access , Ensure the performance requirements of data access through clustered performance integration .

XDFS Networking support from 10GE To 200GE, or 56G To 200G Of InfiniBand The Internet , The front and rear ends do not limit the networking mode , To maximize the network access ability to meet the actual environment of users .

Storage media supports SSD as well as HDD, And support data layering , Select the corresponding storage medium according to the data heat , And automatic circulation between different media , To meet the access needs of different business systems . And the final data can be archived to Blu ray media or public cloud through the embedded cloud archiving function , To meet the requirements of long-term storage .

The project design

Scheme topology

Networking instructions
Design three independent networks , Respectively ： Public data network 、 Data private network and management network .

（1） Public data network

Used to carry image data to XDFS Storage pool upload 、 download 、 Query and other operations on bandwidth 、 Time delay 、IO Response, etc. have great performance pressure , Use a separate 10GE Ethernet link ensures transmission performance .

（2） Data private network

XDFS Balance of data within the cluster 、 distribution 、 Interaction 、 Verification requires the support of data private network , Also for bandwidth 、 Time delay 、IO Response, etc. have great performance pressure , See also the description of data public network .

（3） Management network

The management network is used to log in XDFS Clustered WEB Management interface , Complete the configuration 、 management 、 Status query and other operations , The network has little performance pressure , Connect to a Gigabit switch , It can also be reused with the user's existing network .

（4） File archiving network

Blu ray disc library and XDFS The data public network connection of the cluster , Link in 10GE Switch , The archiving and fetching links of data are transmitted through the public data network .

Redundancy mechanism and capacity

XDFS Support multiple copies and N+M Erasure code （EC）, Provide hard disk level 、 Node level 、 Cabinet level and other multi-level fault domain protection capabilities .

It is recommended to adopt 3 node XDFS colony ,2+1 Erasure code provides data redundancy protection , Up to 1/3 In case of node failure , Data is not lost , Mating hardware RAID Mechanism , Form a hard disk level 、 Node level 、 Cabinet level and other multi-level fault domain protection capabilities , Improve the overall reliability of the media resource system .
The available capacity is calculated as follows ：

1	The nominal factory capacity of the hard disk is 1000 carry , Binary is 1024 carry , Thus every TB There is an appointment 91% The capacity difference of ;
2	2+1 The space utilization rate of erasure code is about 2/3;
3	RAID5 The space utilization rate is about 92%.

（ From the above ：10TB * 24 slice * 5 node * 91% * 2/3 * 92% ≈ 400TB）

File archiving and migration instructions

Blu ray disc library communicates with XDFS connected ,XDFS Files can be automatically migrated and archived into the Blu ray library according to the set policy . The client can see the pointer file of the migrated file （KB level ）, This pointer file indicates the archive information of the migrated file , When the user calls the file ,XDFS According to the information of the pointer file , Send file call command to Blu ray system , Fetch the file to XDFS The original address , For user's use .

Allow users to set migration policies , By capacity 、 Visit Frequency 、 The last access time and other thresholds perform data migration and archiving , The archived data passes API The interface is pushed to the Blu ray server , Perform the write operation of data to optical media . There is no need to adopt another management software in the whole process , It can be completed without human intervention .

03 Program advantages

● Build a unified image lifecycle management platform for users ;

● Establish a unified image content storage and access platform , It can support the unified processing of activity data and static historical data of different applications ;

● Provide appropriate interfaces , It is convenient for related applications to call

● Establish storage and archiving mechanisms for different applications , Online storage according to policy 、 Realize automatic data migration and callback between archive storage , It can support massive data archiving , It can meet the storage time requirements of different applications ;

● Provide a wide range of access protocols 、API And other interface schemes , It is convenient for all business systems to call 、 load 、 Show and use images to conduct business efficiently ;

● Establish a future oriented full content management platform for the bank , Support vertical 、 Horizontal scaling , Meet future business types 、 The increasing demand for business volume and various data content generated by business .

04 Application scenarios

Finance at all levels 、 negotiable securities 、 Various image business application scenarios for insurance users , Such as ：
1、 Bank double entry 、PDF file 、 Historical material scanning
2、 Securities trading authorization contract filing 、 Filing of transaction certificate
3、 Insurance claim documents are saved 、 Insurance contract filing
4、……

The transformation of financial imaging platform is based on extreme digitalization , Based on various unstructured business data , utilize AI、 Big data and other new technologies , Provide more efficient 、 More personalized financial services are the core of this change . In recent years ,TaoCloud Has always been a trusted partner in the financial services industry , With our years of professional experience , It provides professional technical support for the digital transformation of various financial institutions , Participated in and led the transformation projects of some large financial institutions .

Intelligent, stable and efficient 、 trustworthy