Data Storage System Design
As business develops, the company ' s number of file servers increased significantly, with disk terabites (Pic. 1), resulting in poor and difficult infrastructure management. All applications in the company and file servers are directly connected to disk systems (DAS connectivity), which in turn results in low repository productivity, overloading at peak loads, costs associated with additional space for equipment, and high electricity and cooling costs.
At the same time, if additional volume or productivity increases are needed, new applications or file servers should be added, thus further complicating infrastructure management. The company currently has 5TB data distributed between file servers, servers exchange and two SQL servers. Each server contains about 500HB data on average. Each file system is designed with a 100 per cent growth reserve and is deployed on the internal SATA with 1TB volume disks.
Problems encountered by the company:
Large volumes of unused disc space
The company currently loses 50 per cent of the total disc volume (5TB data stored on 10TB disk space). This results in proportional losses on electricity and cooling in the server. The current disc space is also lost in counting the storage of multiple copies of identical files.
The location for future growth cannot be redistributed between systems
The divergence space allocated for further growth is evenly distributed among all systems and cannot be redistributed between systems if the real growth in the data volume does not match the expected. 100 per cent of the reserve comes from the worst option, as it is difficult to predict precisely the level of growth. While some data on some systems are growing relatively slowly, others will experience strong growth; and unfortunately, the low-growth location cannot be redistributed for rapid growth systems.