| Service and Solution → Centers of data analysis → Data storage systems |
In order to arrange control function and minimize costs, enterprises and organization are trying to optimize existing business-processes for productivity, accessibility and scale. At this moment information becomes the most valuable asset, and its accessibility and topicality is a key factor to success in competitive environment. Such situation stimulates wide scale development of data storing technologies.
Modern storage system, while providing reliable storage of information resources, must provide operative access to them. Such systems are built on the basis of disk arrays, accessible infrastructure to them, software and hardware complexes for control of data storage and reserve/archive systems.
Network data storage.
Infrastructure for access to data storage equipment is responsible for operative provision of information to application and end-users. Thus this subsystem develops more dynamically: from ordinal direct connection of disk arrays to servers (DAS – Direct Attached Storage) to widely spreading technology with the use of network in a role of infrastructure for data access. SAN (Storage Area Network) Technology uses dedicated network based on the Fibre Chanel protocol and complete data exchange on a block level. NAS (Network Attached Storage) devices, in fact it is specialized file-servers, use IP-networks as a transport, and operate on a file level. The optimal option is using SAN as consolidated infrastructure of data storage, including NAS devices. With this option the access type is identified basing on concrete requirements of the customer
The main advantages of network data storage are:
- High scale. As the volume of data grows the scale of solution becomes more important, also decreasing cost of data storage.
- High performance. Using Fiber Chanel protocol for SAN (Gigabit Ethernet for NAS) provides high speed of data exchange.
- High accessibility and reliability. Because of module architecture it becomes possible hot reserving of component storage systems and elimination of spate point of failure
- Effective realization of data migration procedures (reserve copying, replication); reduction of workload on local area network (LAN) and computing complexes, thanks to separate SAN network storage and hardware means for data migration
- Performance of data transmission on big distances and possibility for creation distributed CDP.
- Flexible possibilities and redistribution of disk memory and consolidation of data storage devices.
- Realization of safe access to data from heterogeneous environment
With increase in volume of data usage of SAN is becoming more justified solution from economical point of view. Thanks to wide adaptive features of the network data storage, integration of these technologies gives possibility to forecast risks and periods of services idling. This provides high accessibility to any type of data, what decreases costs of data storage considerably.
Integration of networked data storage is one of the most perspective direction for expansion corporate information network systems. Our partner in building SAN are companies Brocade Communication Systems, EMC2, HP.
Data replication
Data replication can help to solve one the most actual tasks:
- Data reserving – safe storage of actual copy of the information on alternative devices of storage or remote platform
- Data consolidation – data copying from remote to central node.
- Control over local data copies for providing information support for the whole organization.
Such operation can be provided by internal program complexes means. However in this case integration of centralized process of replication is impossible. At that data replication on the application level cause extra workload on computing powers of the servers. With the purpose of leveling disadvantages we are using special program replicators or hardware means for data replication.
In spite of way of replication there are two modes that can be applied – synchronized or asynchronized. Synchronized mode of replication provides absolute data actuality, but it requires application of broad bandwidth of communication infrastructure, e.g. integration of SAN.
Data reservation
Effective way for elimination of the information loss, which occurs because of failures, is mirroring. However with this way of data reserving it is impossible to solve quite important issue – wrong action of the user. Reserve copying of the data allows to avoid losses as in case of failure of hardware and software, as in case of operators errors.
During the process of building reserve copying systems and data restore, two major tasks should be resolved:
- Decrease of the “window” of reserve copying (time of application stop – zero downtime)
- Decrease of the data traffic of reserve copying and computing resources cost reduction
With solution of the first task there is possibility to minimize application stop time, using special methods such as immediate copy (snapshot). Using this method the idle can be minimized to seconds. Immediate copy this is snapshot of the condition and recording data location. Using this information data can be restored in case of loss or damage.
The second task can be solved using switching architecture SAN. In this can reserve copying processes and data restore can be organized without LAN resources (out-of-network reserve copying or LAN-free backup) and servers (out-of-servers copying or server-free- backup). Because of centralized scheme of data reserve the number of devices (tape or other) can be reduced.
In case of LAN-free back up data from disk arrays is transported inside SAN to tape drives, which connected to storage network.
In case of sever-free backup data is moved from storage devices to tape drivers witout direct interaction with servers – data exchange occurs between storage devices inside SAN on the level of data blocks. However there is possibility of file-by-file information restore. The main advantage of server-free backup is decreasing workload of server computing resources.
