System Integration Solution

Time: 2016-07-01 15:01:18
Click: 6738

Centralized Backup Project of Xiamen XX Group

Background and Requirements

Current situation and existing problems of the system

In the current backup system of Xiamen XX Group, all the services adopt the backup management called software Symantec backpack exec 12.5 and tape library, partially sharing the storage to backup related data. The business system of each factory is independent of its own backup, lacking centralized, unified and effective backup management mechanism, so that it cannot meet the current requirements of research and development and production on data protection, which undoubtedly increases the risks of data processing, the difficulty of management and the complexity of maintenance.

Low Reliability of File Server

At present, all file servers are deployed on one single device, and some are even on an ordinary desktop device.

The file server is not deployed by double clusters, so that once the server fails, the entire business system will shut down, causing serious problem of service breakdown, which represents the deficiency in the overall reliability.

Storage Limit & No Valid Backup Data(A Quite Dangerous Situation)

Some plant file servers use Symantec backpack exec 12.5 to back up data into the tape drive, but due to storage limit, it is difficult to succeed when backing up unstructured data on the file servers, which leaves a lot of hidden problems to future data restoring.

Incomprehensive Data Protection & Insufficient Cover of Important Business Systems

Only some of the business systems are involved in the backup management system, and other business relies on the manual manipulation of administrators and maintenance staff. Moreover, data protection does not cover all the business systems. Due to the lack of centralized and unified backup management platform, the complexity of system management and maintenance increases, so as the cost of management and maintenance

Low Utilization Rate of Stored Resource & Low Backup Efficiency

On plant file servers we found that a large number of files scarcely related or even irrelevant to work accounts most of the storage space, especially those audio and video files of * .MP3 and *. AVI format, which causes severe waste of storage, increased backup cost and invisible increase of backup window, and regularly fails the current backup system during its running.

1.5 TB data of one plant file server need more than 100 hours to backup, yet the backup is unsuccessful the low efficiency.

Low Efficiency of Management and Maintenance &Increasing Cost

Each plant system is independently planned and constructed. The utilization of resources such as the mainframe and storage are chaotically planned and separately managed. The lack of unified means makes management and maintenance more complex and also increases the cost of maintenance.

Goals of Project Construction

The deployment and construction of centralized date backup and storage aims at the mode of two-dimension centralization and horizontal data distribution and collection.

1.     Rebuild the redeploy each plant file server through clustering and storage mirroring to achieve centralized storage and thus improve the availability and reliability of data.

2.     Back up data locally in the current project and store data to the headquarter in future project. Through the recovery of backup data, the headquarter can access valid data directly such as opening the files.

3.     Provide database support for headquarters to build data platform in the future and deep analysis of relevant data.

4.     Due to the limited network between plant and the headquarter, it is necessary to delete repeated local data in each plant by deduplication and rebuild a new backup network in future data guard for data remote backup.

5.     Meet the demands of cloud calculation, virtualization and oracle database backup in the future.

6.     Achieve unified management to meet the need to construct integrated and centralized data backup and storage.

File Server Cluster Build

Designing Requirements

In order to cope with the low reliability of the system caused by single device operating, we adopt the method of double-machine cluster plus redundant mirroring. The establishment of a secure mainframe and the logical structure of storage prevent data loss and single point failure of the mainframe and storage, ensuring data reliability, availability, integrity and consistency. It also ensures that the file server system provides continuous services to external users and does not lose data within the tolerance.

Designing Topology


Centralized Data Backup and Storage

An Introduction to Program Design

The scheme is based on the backup method of EMC Data domain + NETWORKER, which provides different levels of data backup schemes for major enterprises. It can realize unified management, flexible expansion and cost control, which effectively solves many problems of major enterprises' demands on operating and maintaining data security.


Project Configuration List

Centralized Backup

1.     File Server Cluster System







document server

rack server

4x AMD Opteron 6320 2.8GHz64GB memory, (16x4GB) 1600MT/s
4*300GB 10K RPM 6Gbps SAS 2.5 inches hot-plugging hard disk
RAID controller 512MB cache *1
2* Qlogic (QLE2560) 8Gbps PCI-E FC HBA – single port
2*1100V redundant power supply, 2*5 meters,  multi-model, LC-LC, optical fiber cable
2*1 jumpers, C13/C14, 2 meters, DVD+/-RW ROM, SATA, built-in



EMC optical   fiber storage


2U24 disk, dual controller

8*8Gb FC main channel, dual lithium battery

2*8GB Cache

redundant power supply

4*SFP (8GB), 8*ISCSI port

2.5" SAS hard drive with 900GB port

900GB 10K*24, hard drive with bracket for 2U24 disk;

2*global hot spare

3*raid 5 loss disk



optical fiber   switch


24 port switch, 24 port activation, single power supply (fixed)24*8Gb shortwave SFP, Web tools, Zoning software authorization, no cascade, one-year warranty to the factory



storage fiber   jumper

storage fiber   jumper

connect storage with document server (5M)



twisted pair

twisted pair

6 types of twisted pairs



dual cluster   software


Symantec storage foundation HA