This year is the 50th anniversary of the world's first hard
disk drive storage system-the 350 Disk Storage System and the RAMAC computer
launched by IBM in September 1956. At this juncture, it would be worth looking
at the current challenges for the CIO with respect to Enterprise Storage and
draw relevance to future technology advancement that are expected to occur.
Exponential Data Growth
We are at the start of a worldwide data explosion that will make the booming
population growth of the next 50 years seem paltry by comparison. By 2050, the
world's population is expected to grow from 6.5 bn to 9.1 bn. A more
surprising prediction is that automated monitoring and data collection devices-unheard
of 50 years ago-may soon outnumber people. More people plus more devices plus
more time equals more data. According to IDC's Worldwide Disk Storage Systems
Forecast 2006-2010, we can expect worldwide data to grow at a compound annual
growth rate (CAGR) of 50.6% through this decade. This growth rate means that
every 5.5 years or so the data that the world as a whole needs to manage will
increase by a factor of 10.
It is now easier than ever to create and access data. More than
one billion users are continuously creating Internet data alone worldwide, and
that number increases daily. The proliferation of cell phones, laptop computers,
and PDAs, which connect to data through several access points-private
networks, wireless networks, and the Internet-all contribute to the rapid
accretion of data.
Environmental sensors, video monitoring systems, and
communication recording systems used for organizational continuity and security
create a huge volume of essential data throughout the public and private
sectors.
Unstructured data-the billions of files and emails not stored
in databases-is the greatest contributor to this growth and can represent as
much as 70 to 80% of the data an organization has stored online. Adding to this
category is the daily conversion of extensive paper archives.
Deliberate copies of data files also contribute to the growing
volume of data. Important data is routinely copied to multiple locations to
protect against all types of disaster. In the field of data mining, for example,
huge databases are often copied for the purpose of running business intelligence
queries. Numerous temporary copies of large databases are also created
deliberately in the process of application development and testing. Across the
globe, important data is routinely copied to multiple locations to protect
against all types of loss.
Data Management Costs
As the volume of data grows, so does the complexity and therefore the cost
of managing that data. Specifically, the complexity of data administration
increases logarithmically with the increasing volume of data.
The Business |
Don't Keep: Temporarydata that is deleted by users when no longer needed. Keep Until: Data subject to regulation that will be kept for a defined period before deletion. Keep Indefinitely Because I Don't Know: Data that doesn't get deleted. |
The cost of housing data is a drain. Continuous advances in
storage density make it less costly to save data than to decide what is worth
saving. The possibility that the data could have value to future applications
discourages indiscriminate culling. Due to the continuously declining cost of
raw storage, the peace of mind that comes from keeping all this data available
has been at a price. CIOs and IT administrators have been willing to pay.
Access requirements are more stringent. The demands for access
to data are also becoming more challenging. Globalization has effectively
eliminated windows of opportunity for planned downtime. While New York sleeps,
Shanghai works, and data must be available around the clock. Today's online
world demands 100% data availability.
Legal requirements are more complex. As more and more business
operations are recorded and stored digitally, the thicket of laws and
regulations governing businesses and data becomes denser, and the consequences
for failing to comply with these regulations become more severe. No matter what
the country, the burden of these laws and regulations invariably falls on the IT
department.
Business risks have greater consequences. The global online
economy demands unprecedented data availability and security. In the short term,
a data interruption can have deleterious effects on a company's bottom line.
In the long term a data outage can damage a company's reputation and result in
serious financial effects.
Next Gen A focus on the next generation of storage technologies being developed |
Evolution of Core Unit-HDD The basic component of storage unit-the HDD-has undergone multifold advances in terms of size, data density and cost. So much so, that today's systems have a cost contribution of 10% or less from the HDD. The HDD still appears to have considerable life left in it. Even though a slowing in the rate of progress is projected due to significant challenges, there is a widely held view that no alternative technology is likely to provide serious competition to the HDD in the enterprise for the next ten years. Nevertheless, there is increased interest and activity in alternative storage devices. Presently available alternative device technologies, such as semiconductor memory DRAM (dynamic random-access memory) or Flash, are still about two orders of magnitude more expensive than the HDD, thus ruling them out as storage devices in enterprise storage systems at this point of time. Storage-class Memory A new approach to creating faster storage, this is focused on creating low-cost, high-performance, high-reliability solid-state random-access storage that could compete with or replace disk drives and/or flash memory. Applications of this technology will range from pervasive mobile devices to storage controllers and would possibly include rapid-booting PCs, which could start up in a second or two after power on, not minutes like today's current systems. Intelligent Data Storage Future storage systems will be more than repositories for data. They will also include a wide variety of modern data management and analytic features that will enable more efficient management and utilization of data, which will allow storage systems to help companies with fraud detection and identity recognition. Storage Systems that Compute Smart movement of computing power is enabled by logical partition (LPAR) technology, which allows virtual servers to be created on the storage server. This can accelerate applications by harnessing storage server resources. The intersection Advances in storage management are at a stage where we can say that virtualization and autonomic computing intersect each other. Managing the proliferation of data is becoming a huge and expensive headache for enterprises. Solutions that are based on open standards that use autonomic policy-based protocols to manage heterogeneous storage infrastructures efficiently and economically from a single point of control will gain importance. |
Managing Data
There is already too much data for human management. At the end of the day,
there are just three categories of data: Don't Keep, Keep Until, and Keep
Indefinitely Because I Don't Know. Of these, the Keep Indefinitely is by far
the largest category. Not all data in the Keep Indefinitely category will be
called upon in the future. But considering that one backup tape equals 9,000
tonnes of paper, it is clear that trying to figure out what is worth keeping in
the Keep Indefinitely category would be an overwhelming task. It is also clear
that any culling of data in this category that relies on human intervention will
never get done. Today, saving all data is easier and cheaper than deciding what
to throw away. To plan a storage environment that deals effectively with
requirements for access, security, regulatory compliance, cost control, and
change, IT architects must assume multiple roles. They need to be lawyers,
security guards, financial controllers, diplomats, technologists, and fortune
tellers. Because it is not reasonable to expect their IT architects to be
experts in all these areas, CIOs are now turning to their storage vendors for
help.
Storage technology has been continuously evolving to address the
challenges that had been faced by the CIO community. As a result of this
evolution, the vendor community had also been bringing newer and newer solutions
and concepts to the market. The last couple of years have witnessed
Virtualization, ILM, VTL, iSCSI and several other offerings from the vendor
community.
Sanjit Sinha
maildqindia@cybermedia.co.in
The author is General Manager, Research, IDC India