Practice of super integration and transformation of core production business of public funds

author brief introduction ：

Xia Yan   Evergrande Fund   Deputy general manager of information technology department

Graduated from Chittagong computer and communication Huangpu Beiyou , Twenty year finance IT Experience , Now he is the deputy general manager of the Information Technology Department of Yingda fund 、 Information technology consultant of securities and futures industry of China Securities Regulatory Commission , Mainly responsible for the fund enterprise IT Information construction management and planning .

Abstract ：

This article introduces Yingda fund IT The background of the comprehensive transformation from infrastructure to hyper converged architecture 、 Assessment process 、 Experience of deployment scheme and use , Especially for the small-scale computers that originally supported the core production business SAN Storage related concerns and targeted assessment and scheme design , Hope to bring reference to peers in the industry .

One .  Background introduction

Yingda fund was IT The architecture has always adopted the following common patterns in the industry ： Minicomputer processing SAN Store and run the core database ,x86 Server plus SAN Storage runs other businesses ：

chart 1： Yingda fund was IT Architecture diagram

The reason for the initial launch of the super convergence architecture was that the University hoped to make a disaster recovery based on the business needs of policy supervision , And identified the following core requirements ：

1. Of the plan RPO I hope that according to the requirements of the regulatory authorities, it is lower than 5 minute ;
2. Limited budget , Hope for the most cost-effective solution ;
3. because IT The staff of the Department is limited , Therefore, we hope that the disaster recovery scheme of the data center can be easily maintained ;
4. The infrastructure performance of the production and disaster recovery center must support the daily business pressure of the core database system , And disaster recovery needs to have the same business processing capacity ;
5. The disaster recovery room is planned to be rented IDC Computer room , Space is limited , I hope the less equipment, the better , At the same time, we should ensure the convenience of subsequent expansion ;

According to the traditional disaster recovery center scheme , Need to buy another set of storage , And purchase the replication software that comes with the storage , Two stores of the same brand and model are needed at the same time , The overall plan is more expensive , And it's open , Poor maintenance and scalability . in addition , at that time IT There are other problems with the infrastructure , for example ：Oracle RAC It's also deployed in IBM On a minicomputer , Old equipment , Consider equipment risk and complex operation and maintenance issues , Hope to eliminate as soon as possible IBM Small computer .

Because I heard about distributed storage with super converged architecture many years ago , And think that this kind of program may become a trend , Will gradually replace the traditional architecture , Now if you spend a lot of money on the traditional architecture, it may soon be eliminated , So we found SmartX、 Lutanli and other manufacturers come to introduce . After communication, we found that this architecture has many advantages , And the maturity of the product at that time , The coverage of application scenarios has been greatly improved , So we started to evaluate whether the overall architecture can be upgraded to a hyper converged architecture , And based on this architecture to build a city disaster recovery .

Two 、 Evaluation and design of super convergence architecture scheme

Compared to the traditional solution , The advantage of super convergence architecture is still very obvious , for example ：

1. be based on x86 The entire server IT Both architecture and disaster recovery architecture can be simplified , The reliability of the whole system can be further improved ;

2. Better scalability ,IT The team can meet the needs of business resource delivery more quickly ;

3. Total cost of ownership can be effectively reduced .

Although there are the above advantages of super Integration , But because it's about the company as a whole IT Basic transformation , So the team still has concerns about reliability , Can the super fusion architecture based on distributed storage and virtualization really replace the original small computer plus traditional storage ？ After all, there are not many cases of business transfer from small computers to super converged architectures .

In response to the above concerns , We have carried out scheme evaluation and design consideration from several directions ：

1. I visited many fund companies, insurance companies and other customers to understand the application situation , We know that the application of super fusion in production system is very common ;

1. Through the analysis of super fusion products POC, In terms of product reliability 、 Fault tolerance 、 performance 、 The function has been fully understood and verified .
2. Make full use of existing equipment combined with Oracle DG Technology has done data synchronization at the database level , That is, for the core business , Make a double insurance on the basis of hyper converged distributed storage replica .
3. Use virtualization based CDP Products cooperate with super fusion to build disaster recovery center , Do three insurance .

The final scheme structure is as follows ：

chart 2： Yingda fund is now IT Architecture diagram

3、 ... and 、 Super fusion architecture application effect

At present, all the production systems of Yingda fund have moved to the super integrated platform , Include O32 System 、 Valuation system 、TA System 、 Direct selling system 、 Various types Oracle database ( investment 、TA、 The valuation 、 Risk control 、 Data Center ）、 The market 、 Offer, etc , In the past year, the application effect of super fusion transformation is as follows ：

1. The system has good performance , It has been running steadily for more than one year without any problems , The new architecture based on super convergence can completely carry the original IBM All kinds of services, including those on minicomputers ;
2. The whole data center has only standard x86 Servers and Ethernet , And distributed storage in the fault automatic recovery , It's very intelligent in terms of capacity expansion, automatic data equalization, etc , Operation and maintenance personnel can get rid of the tedious and inefficient operation and maintenance, and carry out more innovative work such as new technology research ;
3. Cabinet space is saved 83%, Using super fusion high density all in one machine , Previously, all hardware devices used in the production infrastructure 12 The number of cabinets is reduced to 2 Cabinets ;

chart 3： Comparison of cabinet space before and after the upgrade of Yingda fund data center

1. The entire data center network is also greatly simplified , Here is the network architecture before and after the upgrade ：

chart 4： Yingda fund data center network architecture before upgrading

chart 5： The network architecture of the upgraded data center of Yingda Fund

1. Get through the two-tier architecture of production cluster and disaster recovery cluster network , adopt CDP Tool implementation RPO Second level 、RTO Minute disaster recovery capability , Virtual machines on two clusters backup each other , One button switching , Greatly reduce the complexity of disaster drill , This can be introduced later ;
2. The scalability of the system is very good , Fast online , These agility IT There are several recent examples of the benefits of . The first example is that companies are anxious to start quantitative investment testing , We need to build a test environment as quickly as possible , And quantization requires high-performance servers , Because we use a new architecture like hyperfusion , Soon supported the company's quantitative business . Another example , It's the company's recent rush to launch a new business system , Built test environment , It's like copying the production environment , Based on super fusion architecture , Eight or nine virtual machines were in place very quickly .

Four 、 Summary and experience

Through more than a year of stable operation in the company , It proves that super convergence can carry all kinds of business including core production , It's easy to maintain 、 Easy to expand 、 Agile features such as fast launch also meet the needs of the company's financial technology transformation , The system has been expanded recently , Will continue to support the company's business development .

The last thing I want to say is , Although the super convergence architecture is relatively mature , But the situation of product realization and service varies greatly from manufacturer to manufacturer , We must choose a professional product manufacturer who is really competitive in super integrated products , For example, abroad is mainly Nutanix, and VMware, In China, it is mainly SmartX, Foreign product brands are definitely stronger , But the price is higher , Local service response may not be timely enough , After comprehensive consideration, we finally chose domestic professional manufacturers SmartX, Relatively speaking, the price is more reasonable , Services are also more responsive , So far, the results are satisfactory .