All-Flash Storage Reduces Transaction Waiting Time for Nanhua Futures

Many stock investors are familiar with the frustration of transaction system login failure and transaction failure. In fact, transaction failure and system breakdowns are not rare occurrences. For stock investors, such incidents are unacceptable, because significant amounts of investment may be lost in a matter of seconds.

According to technical engineers of securities companies, system breakdowns result primarily from extreme conditions not being fully considered in system design. The preset threshold of transactions is low, and technical personnel do not respond quickly enough when emergencies occur.

We are going to discuss how core transaction platforms in the financial industry can be both fast and stable to improve customer satisfaction.

Founded in 1996 and owning 15 billion RMB in total assets, Nanhua Futures engages in commodity futures brokerage, financial futures brokerage, futures investment consulting, asset management, and securities investment fund agency sales businesses. It is among the top 10 futures companies in China in terms of business scale. Its futures trading volume, total deposit, and asset management business scale are also leading in China.

New Requirements for the IT System from Growing Business

In the futures industry, time is money. Profits may increase or decrease by billions in one second. As a leading futures trading company, Nanhua Futures has attached great importance to its IT system. It looks to build an open, fast, stable, and secure futures trading and settlement platform to support the rapid development of futures business. Its data centers in Shanghai and Hangzhou provide the IT environment required by the Shanghai Futures Comprehensive Transaction System (CTP). The two data centers adopt parallel deployment; each is capable of supporting all business of the entire company. In addition, Nanhua Futures has leased VIP equipment rooms and cabinets from various industry data centers and carriers to deploy its transaction system: Dalian Commodity Exchange, Zhengzhou Commodity Exchange, Shenzhen Stock Exchange, China Unicom Jiaxing, and PCCW in Hong Kong. The purpose is to provide a direct and fast transaction channel with an IT system latency of less than 1 ms. To prepare for future business growth, Nanhua Futures wants to upgrade its existing IT infrastructure.

At the same time, stability and accuracy of the market, transaction, and settlement system are the most important requirements for the IT system of futures company. Failure of the IT platform of the transaction system may cause service interruptions and massive losses. Therefore, the core goal is to improve the reliability of the core transaction platform and ensure 24/7 availability of the transaction system.

“2 Cities, 3 DCs, and 4 Copies”: Ensuring Reliability of the Core Transaction Platform

There are two DR protection methods in the industry: the gateway-based active-active method and the gateway-free active-active method. In the first method, a gateway is introduced into the existing network to improve reliability. However, this adds a potential point of failure into the network, makes the networking more complex, and presents the risk of misoperation. Furthermore, adding the gateway increases system latency and reduces system performance. For that reason, Huawei uses the gateway-free active-active method in its transaction platform solution, HyperMetro. Load is balanced between the active-active sites, and failover will not cause service interruption, achieving RPO=0 and RTO≈0. The network does not become more complex, because no gateway is required. The active-active solution can be smoothly upgraded to a gateway-free 3-DC solution, which provides 99.9999% availability at the solution level, preventing breakdowns from occurring on core applications.

After several rounds of communication and solution verification, Nanhua Futures chose Huawei to upgrade its IT system. Huawei deployed OceanStor Dorado all-flash storage to support the Shanghai Futures CTP. Two sets of Dorado were deployed in the Shanghai data center A to ensure high availability of mission-critical services. One Dorado each was deployed in Shanghai data center B and Hangzhou data center C. Service data of data center A is synchronized to data centers B and C, ensuring service continuity in case any data center fails. The innovative 3-DC solution with four copies provides sound data protection for Nanhua Futures.

Networking

All-Flash Storage Reduces Transaction Waiting Time for Nanhua Futures

Configurations

Data Center Device Model Device Configurations Device Quantity
Production center OceanStor Dorado5000 Dual controllers, 16 x 16Gb FC, 20 x 1.8 TB SSDs, snapshot, inline deduplication, inline compression, active-active storage, and remote replication 2
Disaster recovery center OceanStor Dorado5000 Dual controllers, 16 x 16Gb FC, 20 x 1.8 TB SSDs, snapshot, inline deduplication, inline compression, and remote replication 2

Efficient IT System Supports Long-Term Business Development

Previously, the core equipment room of the primary transaction system stored data on traditional HDD storage. 13U of cabinet space was required to meet capacity and performance requirements. Huawei OceanStor Dorado all-flash storage (inline compression and deduplication enabled) needs only 2U of cabinet space. This means a reduction of 85% in space occupation, 71% in power consumption, and 70% in OPEX over the course of 3 years.

After deduplication and compression are enabled, Huawei’s all-flash storage system can maintain stable performance and quick response to transactions during peak hours. Incremental backup from the Shanghai data center to the Hangzhou data center takes only 15 minutes, in contrast to the 24 minutes required when HDD storage is used. This agile and efficient IT system helps Nanhua Futures cope with the rapid growth of services as well as long-term development.

All-Flash Storage Reduces Transaction Waiting Time for Nanhua Futures

Smooth Data Migration Makes Data Center Evolution Easier

Data migration is a headache for everyone. Migrating data between different operating systems, application software, and device models presents a major challenge to compatibility. Additionally, service interruption during migration is best avoided. The core transaction platform of Nanhua Futures has over 20 service systems and carries data of different models of databases and VMs. In addition, it involves compatibility certification of 100+ types of middleware. All of this makes the platform migration extremely challenging. Futures business is very sensitive to interruption time and migration duration. Any service interruption during migration will lead to customer complaints and huge economic losses. Therefore, Nanhua Futures had high requirements for the migration quality. Facing these requirements and a tight schedule, many vendors gave up on the project. However, dealing with complex situations is what Huawei does best. Huawei provides perception-free data migration service based on the heterogeneous virtualization migration solution. Storage on the live network was seamlessly taken over. Around 10 TB data of the core transaction system was migrated in 4 non-transaction hours. Production was not affected in the process, and test personnel did not perceive any impact from data migration. This smooth and stable experience earned Nanhua Futures highly positive feedback from customers.

Nanhua Futures moved its core transaction platform to Huawei OceanStor Dorado in November 2017. Since then, Dorado has been running stably for seven months. The IT system not only meets the requirements of the futures system for fast transaction response, but also reduces maintenance costs. The innovative “2 cities, 3 DCs, and 4 copies” solution based on all-flash storage ensures sustainable and stable business development.

(Contributed by Wang Jiaxin, IT Product Line)

The post All-Flash Storage Reduces Transaction Waiting Time for Nanhua Futures appeared first on Huawei Enterprise Blog.

Source: Huawei Enterprise Blog