Background

As the growing popularity of electronic trading, process-centric banks, and centralized service handling continues, unstructured data, including check images and audio/video recordings, are widely applied in the finance sector. IDC, a consulting firm, reveals that the big data era of the finance sector is around the corner, and that 80% of service data will be unstructured data in varied formats, like images and audio/video files. From 2010 to 2020, unstructured data is skyrocketing by 44 times and urgently calls for data management reforms. Massive unstructured finance data is generated primarily because:

• The construction of process-centric banks and service handling centers requires frequent and a huge amount of check image file exchanging and storage among front-end, mid-end and back-end transactions.
• Electronic financial services, such as call center and mobile marketing, demand audio/video recordings for proof.
• The post supervision center scans a large number of printed checks for digital, ICT-enabled, and network-based service supervision.

According to statistics of a China's state-owned bank, its over 1.8 billion online check images scattered in its over 300 branches exceeds 150 TB. And this data increases by daily 1.8 TB that is generated by over 20 million transactions per day. Such massive unstructured data poses many challenges for financial organizations in terms of its data management and maintenance.

Customer Challenges

The unstructured data has the following characteristics:

• Small files

Almost all files are smaller than 1024 KB.



• A large number of files

Millions even billions of files demands exchange and storage.



• Few access

Most of files are left unvisited within 3 months after the generation.



• Quick search

Service personnel need to obtain demanded documents within the shortest-possible time under varied scenarios, for example, customer complaints, judicial investigation, and service handling.




Massive unstructured data increases the IT management difficulties on the following aspects:



• High capacity expansion costs

According to the industry regulations, audio/video recordings and check image files must be available within a long time even permanently, which calls for large-capacity storage systems. At present, many financial organizations store most of the unstructured files in storage systems that are deployed to store production resources, increasing storage costs and leading to frequent, expensive capacity expansion.



• Backup modes

Some financial organizations do not have professional backup systems and store all data in the production systems. Therefore, once the production storage systems are faulty, it is almost impossible to recover the critical data. In addition, traditional backup software developed to store structured or small amount of unstructured data is found powerless and inefficient when massive unstructured data requests efficient, reliable backup and archive.



• Search speed

To back up and archive massive unstructured data, application developers provide data management tools that can write call center audio recordings and check images into disks/tapes. However, using these tools, service personnel have to manually search databases, load disks/tapes, and recover data, which dramatically decreases the search efficiency and wastes labor resources.

Financial organizations are forced to deploy a comprehensive data protection solution to improve operation efficiency and comply with industry regulations.

Figure 1-1 Data backup requirements for call centers and check images

The planned data management solution must address the following requirements:

• Automatic archive that complies with industry regulations to release more space for production storage and cut down on storage costs
• Automatic backup to quickly recover systems or files in case of system faults or file damages
• Automatic file callback and archive
• Extended storage time and expanded storage space at the minimal storage costs

Huawei Solution

Figure 1-2 shows the architecture of Huawei's Call Center and Check Image Data Management Solution.

Figure 1-2 Solution architecture

Huawei's solution uses an integrated design and offers automatic management throughout the entire data lifecycle. As an ideal solution for a rich set of application scenarios, including unified data backup, archive, analysis, and search, and duplicate data deletion, Huawei's solution is differentiated in ease-of-backup, quick search, and low costs

Ease-of-backup, enabling TB data backup within one hour, improving backup efficiency by up to 10 times, and backing up billions of files



• Uses multiple advanced technologies, for example, Data Classification Enable (DCE), snapshot backup, and multi-stream backup, to back up TB data within one hour and slash backup time to hours from days.

High searching speed, enabling service systems to access archived data without visiting production storage and achieving searching in seconds



• Uses a unique check stub design and enables the system to quickly call back archived files using the file location information stored in the data management system.
• Adopts a novel 2-level index structure and sharply reduces the time of locating the targeted file among millions of files to seconds from over 40 minutes.

Low costs, replacing the production storage with the nearline storage, averting unnecessary production storage expansions, and reducing up to 80% reduced storage costs



• Uses the check stub design to replace the production storage with the inexpensive nearline storage and enable application systems to quickly access the archived files without compromising the operating efficiency of application systems.
• Ensures high data backup, archive, and management efficiency without procuring additional management systems.

Customer Benefits

Huawei's Call Center and Check Image Data Management Solution bring customers the following benefits:

Comprehensive protection for massive data and improved backup efficiency by 10 times



• Provides backup for millions of files and enables 1-hour TB data backup.

Reduced storage costs and no need for production storage expansion



• Provides archive and quick callback functions to release more production storage space that is occupied by the not frequently visited files.
• Replaces the production storage with inexpensive nearline storage to save customer investment.

Automatic, second-level file querying and archiving and imperceptible data retrieval



• Enables application systems to directly retrieve archived data without offline media searching and obtain data in seconds under customer complaints, judicial investigation, and service handling scenarios.