Emerging Enterprise Forum
Managing the data explosion
Enterprise data stores keep growing, and it is critical to
put management processes in place to keep information in rein, says Soutiman
Executive Director, IT Seagate
To satisfy the access and availability demands of todays distributed
and demanding business environments, enterprise data storage is becoming network-centric.
That said, hardware alone is not enough to create an effective, cost-efficient
network storage strategy. Information keeps pouring in from all areas of a business
and leads to data explosion in the enterprise.
Many CIOs continue to buy cheap disk storage to deal with the overload. This
approach will force companies to hire additional personnel and spend more than
necessary to manage their data effectively.
There are many challenges to the enterprise information storage architecture
that lead to the data explosion.
The number of enterprise applications used in organisations is increasing, and
the number of core functions is growing as well. The increase in data volumes
is proportional to the increasing number of application deployments such as
enterprise resource planning, enterprise data warehouses, business intelligence,
data mining, and active data warehouses. Performance and data intensive applications
such as document or content management systems and image processing further
exacerbate this problem.
The performance of the storage system reduces with the rise in data volumes
and creates scalability challenges for the enterprise, says Madan Mishra,
Executive Director, IT, Seagate.
Lack of consolidation is a formidable challenge. In a distributed storage infrastructure,
usage is always poor and therefore the enterprise suffers from data being unavailable
when load peaks. This also creates downtime every time new capacity needs to
There is usually a limited window of time for backup. Chances of data replication
in the various areas of the business are high, and the cost of storage administration
rises in such a situation.
Problems In The Long Run
There may be issues related to security and mobility. With the increased number
of mobile access devices used by mobile workers in organisations, there is a
lot of information flowing through channels previously undiscovered by the system.
There may be data loss which could create legal and compliance issues, and may
ultimately erode the organisations competitive value.
Management of data throughout its lifecycle, including backups, retention and
archiving, becomes increasingly difficult. There are also technology issues
such as input/output performance, bandwidth, bus speed, and less fault tolerance.
Business continuitydisaster recovery and data mirroringbecomes more
Addressing The Challenges
CIOs have to initiate strategies and processes to address all the storage management
challenges created by the data explosion.
CIOs must respond and adapt dynamically to changing business and application
requirements. They must be in step with the evolving needs of the business and
the market, says Mishra.
They must be in constant communication with business heads and try to understand
what the business needs from its information management systems. The storage
infrastructure must have more resilience than that achievable by traditional
A few good ways to simplify storage management
are to efficiently utilise disk space, simplify allocation of disk space
to servers, and allow changes at a physical level with no server impact,
suggests Madan Mishra of Seagate
Disaster recovery measures must be proactive. They should
be flexible and transparent to deploy, scale, and manage. Storage resources
should be configured in such a way that they can be easily shared across servers
and applications in a simplified manner.
A few good ways to simplify storage management are to efficiently utilise
disk space, simplify allocation of disk space to servers, and allow changes
at a physical level with no server impact, suggests Mishra.
These efforts will ultimately improve the efficiency of the overall IT environment,
and provide a higher level of fault-tolerance and resilience.
A few technologies promise to provide more efficiency in storage systems in
future. They are 3 Gbps SATA and Serial Attached SCSI (SAS).
A SATA 3 Gbps interface permits data transfer at 300 Mbps. It can be used anywhere
in an organisation that uses SATA, such as in PCs, notebooks, servers and external
3 Gbps SATA enables high-level performance while maintaining desktop cost structures.
It also facilitates bandwidth aggregation for multiple devices, enabling maximum
throughput. It is fully compatible with 1.5 Gbps SATA and needs no new driver
or cable upgrades.
It is important to note that a 3 Gbps interface does not necessarily mean
higher performance. A 3 Gbps performance will be best felt in multi-drive systems.
The internal transfer rates will still be below 100 Mbps, cautions Mishra.
SAS is the new enterprise interface standard targeted at OEMs
and system builders. It supersedes parallel SCSI, which for over two decades
has been the enterprise interface standard for storage entities such as DAS,
entry-level and mid-range network storage systems.
Here are a few suggestions
for building a cost-effective and high-performance storage infrastructure
according to the type of applications in the enterprise.
- For an OLTP application use high-availability
systems with low I/O rates.
- For a DSS (Decision Support System) application
use storage systems with high I/O rates.
- For information archival use servers with
large capacity but low I/O rates.
- For online backup and mirroring use systems
with very high I/O rates, scope for path failover, and load balancing