by July 16, 2006 0 comments



Once upon a time, the organizations had plenty of time and capacity or face
it that the needs were limited. Backups could be done at night when the systems
were not loaded or were shut down. But since then, the needs have crossed the
roof, systems work around the clock and so do the users, and there’s no period
when you can shut everything down for backup. Today, an organization needs to
store all kinds of data. There are different time periods for which you want to
store this data based on its utility to the organization and regulatory needs.
And then you have storage devices lying around that’ve been exhausted or are
soon going to be all filled up.

All this makes it necessary that there be better storage devices that are
more capacious and faster. For this, researchers have been continuously
exploring new technologies or improving existing ones to help you store a larger
amount of data, for a longer period of time. And all this in a manner that you
do not have to pay exhorbitantly compared to what you do today. The new
technologies should be such that you have to do minimum change in the existing
infrastructure. Some of these include virtual tapes, 4GFC, Blu-ray, HD DVD,
serverless backup, perpendicular recording, PSAN, 3D and online storage that are
standing the wings to join the bandwagon.We take you through the technology
behind each and promises they make for the future.

Perpendicular recording
Let us get down to the
basics. More things can be packed in a space when they stand up rather
than when they are lying down. And that’s the reason why perpendicular
recording is in vogue. Perpendicular recording aligns the bits
perpendicular to the disk, which allows you to fit in more data into the
same space, at higher recording densities. This technology is expected to
increase the capacity of hard drives as much as 10 times. It is the
saturation in the longitudinal recording capacities that has led to this
innovation to counter superparamagnetism. Superpara-magnetism is a
phenomenon that causes interference amongst the magnetic grains that make
up data bits when they are packed beyond a certain density on the disk. As
a result, the grains lose their orientation leading to corruption of data
on the storage device. Seagate has been the pioneer in using this
technology by launching its 2.5 inch, 160 GB drive for notebooks. The
latest we’ve received for evaluation is the Seagate Barracuda drive with
a whopping 750 GB capacity. Hitachi is also backing this technology. At
this speed, we will soon have 3.5 inch disks storing terabytes of data.

Virtual tape library
Tape has always been the preferred medium of backing up data permanently and its
benefits have been numerous. For instance, with tape you get to store data
off-site for a longer period of time that you might need it for. Not only that
you can restore all data lost or deleted from the machines in use. This is
contrary to how the hard-drive storage works. Here the data stored, used as well
as unused, all lies on the machine you are using. While this simplifies the way
you can access the data (tape stores the data sequentially while the disk does
it in random fashion), it is a costlier option compared to the tape-based
storage, both in long as well as short run.

But that’s only one side of the coin. Disk-based storage hasn’t lost the
battle for the advantages it has over tapes. It is faster, and is easier to
handle when you want to work with and restore data. To add to it, disks offer
online storage unlike tapes that need to be kept separately once the data is
backed up on them.

In the light of advantages and disadvantages that each one has, researchers
had been struggling to find out a workaround. And they have succeeded in having
the cake and eating it too-with virtual tapes. Virtual tapes combine the power
of both the tapes and disks, in a manner that lets you reap the benefits of
both. Virtual tapes emulate (behave, look and feel) like a tape, though they are
actually disks. In action, the back-up software sees the VTLs (Virtual Tape
Libraries) as physical tape libraries to perform backup operations. As a result,
while the data is written in a sequential format (as in tapes) on the disks that
emulate as tapes, it is backed up almost 10 or more times faster than the real
tape. Also, you are saved from the hassles of rotating, handling and storing
tapes.

It requires almost negligible or no changes to the back-up processes,
schedules and workflows. While it still doesn’t solve the inherent problems of
storing the data offsite to be used in case a disaster strikes, VTLs can be used
as faster alternatives to store less-used data that can finally be moved to real
tapes whenever needed.

FC and iSCSI
iSCSI, for a long time, was perceived as a better, low-cost storage
networking protocol compared to the Fiber Channel. But today iSCSI has emerged
as an entry level for SAN interface. On the other hand, the FC (Fiber Channel)
is apt for operations that involve connecting servers to shared storage devices.
Most iSCSI based storage area networks use 1 GB Ethernet unlike the FC that has
touched 4G capabilities. The 4GFC products are well suited for high-bandwidth
applications such as voluminous data backups, streaming media and transactional
processing.
Either technology is well suitd for specific needs. To understand this, if a
less amount of data is to be processed, both technologies serve the purpose (but
FC loses out to iSCSI due to its price and difficult deployment process here).
On the other hand, if more data is to be handled, iSCSI’s speeds create
bottlenecks.

Serverless backup  
When a backup is in progress,
your servers-the source and the backup servers-are the most stressed.
During this time they consume a significant amount of CPU cycles and
memory.

The solution
A good option is to offload these servers of this extra load during
backups. The process is referred to as serverless backup, and is a SAN
based solution. This is done when you use a data-mover device that
initializes the back-up process to copy the data directly from the source
(say, a hard disk) to the destination device (back-up devices). The data
mover may be implemented in a switch or a router. There are products
available in the market that have the serverless back-up option that
leverage SNIA’s Extended SCSI Copy command to significantly enhance
performance by taking the backup load off servers and freeing CPU time for
other mission critical tasks.

When to consider
While there are such high-end backup solutions offered by companies like
Legato, Veritas, Computer Associates and IBM, it is not always necessary
that you have to switch to the new ways. Instead, your decision should be
based on your priorities, peak hours and type of data that you need to
back up.

For instance, if you need to backup the
mailboxes of all the employees across the organization, you need not do
that in office hours when the servers are loaded with other processes.

You may go for such solutions if your
business needs include:

  • Frequent backup and restoration of
    huge amounts of data
  • Backup during office hours and,
  • Your priority is server performance

3D storage
Imagine writing data across a volume of space rather than on a two-dimensional
media. This means the same area can be used to store more data. This is the idea
behind 3D storage. Here, you are not only using the x and y dimensions but also
z-axis to store data.

A hologram is formed on a photo-sensitive recording media at a place where
the reference beam and the signal beam carrying data intersect. Holography has
the potential for fast transfer rates and fast random access times. This is
because the data are stored and recovered in parallel-typically 1 million bits
at a time.

Experts predict that in a few years we’ll eventually be able to store
terabytes of information in a space no larger than several CDs stacked on top of
each other. That’s the promise of holographic storage. For more on how it works,
refer to Enterprise Storage Technologies: the way ahead, in our March 2006
issue.

IBM’s Venom for Viper
Venom is the new compression technology that IBM has unveiled. This is being used by the DB2 9 (formally known as Viper) data server which the company has pitted in against other database market leaders as well as storage giants. It is an RDBMS with XML and storage compression capabilities that works well with all Windows, Linux and UNIX systems by using row compression technique to compress data objects in multi-dimensional clusters. Row compression saves disk/memory space by appreciable amounts. This way the system stores the repeated byte strings from a large table in a dictionary and then replaces those table strings with a 12-bit symbol that represents the actual data stored in the dictionary. As a result, the data server can look at all data in the table than that on a given page. Also the repeating sub-strings within a column can be compressed into a single symbol. This not only saves the disk/memory space but also increase compression rates to a good 80%.

HD DVD vs Blu-ray
While the debate on which one will take the market by storm is quite old now, it
is still hot. If it is only price that will govern who will dominate the market,
then the verdict tilts in favor of HD DVDs. But if the battle is out for higher
storage capacities, it will be Blu Rays’ turn. Both HD DVDs (High Definition
DVDs) and Blu Ray disks use shorter-wavelength blue laser to read and write
data. The two are heavily backed by different industry giants.

Blu-ray is backed by Sony while HD DVD uses another technology called the iHD
that Microsoft and Toshiba have worked on. A dual-layer HD DVD packs around 30
GB data on it as compared to the 50 GB (around one and a half times) by Blu
rays, which is by no means small. We have to watch out for some time before
judging which one wins.

P-SAN
Here is a disk-based backup storage technology with which you no longer need to
use fiber-optic cabling to connect SAN components. P-SAN is a storage system
developed by JMR Electronics with the capability to connect to the network as an
extension of the standard PCI bus on personal computers. It uses a PCI switching
architecture to work as an extension of the JMR’s PCI bus using CA’s
BrightStor ArCserve backup solution. Instead of fiber-optic links, the
technology makes use of IP storage. This means that it uses TCP/IP and standard
Cat 5 cabling to connect to the network. It can scale up to large capacities at
an optimum cost, making it a perfect alternative to tape or other costlier
disk-based systems.

Online storage
This is an interesting option that organizations and individuals can explore
alike and there are already more applications than you might first realize.
There are sites that offer a ID/password protected space on the Internet where
you can store your files. Your space can only be accessed by you or by someone
you want to give access to. Today many such services like Foldershare, Bitvault,
Xdrive and iBackup are already operative that give you both paid and free
storage space online. While they give a small amount of storage space for free,
say 25-50 MB, they have paid services for
organizations that would need larger storage space. Some sites offer true group
collaborative functions as well. This way you can access your data from any
computer if you have Internet access and a browser. Though individuals are
currently using most of this online space to store data, organizations can also
benefit from it once these sites integrate efficient security mechanisms within
their services. Its not over yet. As demands on capacities continue to explode,
new technologies will rise up.

Rinku Tyagi

No Comments so far

Jump into a conversation

No Comments Yet!

You can be the one to start a conversation.

<