From NAS to distributed object storage, GF Securities advanced on the way of IT improvement

GF Securities, founded in September 1991, is one of the first batch comprehensive securities traders in China. Since 1994, it has been among the top ten securities traders in China. So far, by virtue of correct development strategy, normative management and perfect risk control mechanism, GF Securities has developed into one of the securities companies highly influential in the market.

With the rapid increase of the online business, in order to prevent illegal operation, money laundering and other financial risks from the source, financial regulators have increasingly strict requirements for regulation of business handling records (audio and video), original vouchers, account opening information and other documents, which may be a great challenge to the information systems of various securities companies.

In addition, in daily securities business, GF Securities has gradually adopted AI technology in its business system to improve business handling efficiency and reduce human cost. Moreover, its unstructured data has been increasing by 40-50% annually, such as customer ID, remote account opening videos, scanned contract copies, customer service center voice and enterprise related data..

Based on the above-mentioned reasons, GF Securities hopes to establish a set of unified content management and storage platform, so as to realize the unified storage and management of the unstructured data, eliminate the information island, and satisfy the requirements for content mining and utilization of unstructured data in the future. Its core demands are as follows:

1. 1.The storage capacity can be provides elastic storage, scaling up to hundreds of PB to meet the demands for data storage in the next 10 years;

2. 
   

2.There are a great number of manageable files up to 10 billion and many small files will be generated from the business. It is not expected that the business performance will be affected by the increase of files;

3.It can support the mode of dual operation data centers and data read-write can be available in different data centers for different business systems in need of no cross network access;

4.In the future, it can support document processing, including document watermark, format conversion, document identification and other functions, and simplify the processing logic of the business system;

5.Advanced technology, in line with the trend of technology development in the future

GF Securities selected SandStone MOS (Mass Object Storage) to have established a content management and storage platform, and gradually stored the unstructured data of its business system from original NAS, MongoDB and local file systems to SandStone MOS in a unified manner; since the platform was launched in 2017, the investment banking, investment research, one counter pass, audio/video recording, container platform image warehouse, e-log, customer service recording and other systems have been directly stored or archived on SandStone MOS, with the number of stored files close to 1 billion.

1. 1.Centralized data storage has simplified the document system: The document system originally needed to collect data from various types of storage devices, but now can have access to all data only through a set of storage;

2. 
   

2. A set of storage has simplified operation and maintenance: It is unnecessary to manage different types of storage, such as NAS and MongDB, and it can achieve PB-level storage and one management interface;

3.More secure data and more efficient access: By virtue of mutual disaster recovery among multiple centers, the business system can achieve local access through a unified domain name and switch disaster recovery automatically;

4.Built-in storage processing functions have simplified the business: SandStone MOS can automatically transform formats and make watermarks for users’ access to ensure access security and uniformity; in the future, it will also integrate quality inspection, OCR and other functions in the storage.