2010 data protection best practices

Cloud?
Tape?
Hybrid?
2010 Data Protection Best Practices

Backups are a curious paradox, are they not?
By Greg Shields
On one side, they’re arguably the most important task of every
IT professional. Protecting your company’s most critical data
really means protecting your company itself, and therefore your
own livelihood. But on the other side, backups are so often
accomplished using technology that hasn’t evolved much past
the reel-to-reel days. In 2010, most IT organizations still find
themselves clinging desperately to ancient tape-based technol-
ogies; technologies that indeed back up data, but do so slowly,
painfully, and sometimes with catastrophic failure.

Why has this paradox become so ingrained into our industry’s
collective consciousness? Why have backups’ best practices so long
involved tactics that don’t really align with today’s datacenter needs?

One source just might be the technologies themselves.
Sponsored by

www.ironmountain.com/cloudrecovery 1

Thankfully, there is a better way. And that way doesn’t necessarily get rid of your existing
investment in tapes. Rather, it arrives as an augmentation to tapes, giving you all the features
you’ve long desired while retaining the quiet comfort of their physical media.

That better way is called hybrid backups. And to best understand how it will dramatically
improve your data protection, let’s look at some evolutionary changes in backup technology
that you might not know about. Along with those changes, you’ll learn a bit about today’s
best practices in correctly aligning your backup technologies to your restore needs.

D2T - Yesterday’s Approach to Backups
Up first is the Disk-2-Tape story that you already know about. Starting here is important in
drawing a comparison with backups’ most recent evolutionary leaps. Tape backups have served
the IT industry for virtually all of its long history. Getting files onto tape has been a major task of
every IT organization since the introduction of data itself.

Yet tape backups are no longer sufficient for today’s datacenter needs. Think for a minute how
an exclusive focus on tape backups can actually hurt the datacenter:

Data storage continues to grow, as do backup windows. In times not long past, backup
windows weren’t a big problem. Knowing that your backups completed between the hours of
3:00 AM and 5:00 AM was a functional solution because most people simply weren’t working
during those hours. That’s no longer the case today. Users work from home, in different time
zones, and in different parts of the world.


At the same time, the amount of data to be backed up just keeps growing. With it comes a similar increase
in the size of backup windows. As you know, tape backups require windows of time to complete. This is the
case because their operation seeks to capture the changes in data that have occurred over that set period of
time. Often just a day in length, but sometimes even longer, backup windows need a time period when data
remains relatively unchanged so that data can be contiguously copied to tape media.

Today’s datacenter needs, however, are maturing to the point where windows are no longer appropriate.
Knowing that you’ve got a copy of yesterday’s data isn’t good enough. Even a 24-hour loss of data could be
catastrophically costly to your business operations. Instead, you need a constant and continuous copy of
right now’s data, something you simply can’t get with tape.

Tapes are linear-read/linear-write. Using tapes, you can’t get that constant and continuous copy of right
now’s data because data takes time to be written to tape. Tapes write data in a linear fashion, copying files
from disk as “the tape” in that tape passes over its head. While today’s tapes are growing ever faster in how
quickly they can transfer data, their actions still consume time at a linear rate. The more data you’ve got to
back up, the larger your backup window must be.

The linear-read/linear-write format of tapes also makes them highly inefficient in backing up data that
changes only slightly from window to window. For example, a file that is backed up to tape and then later
modified must be backed up again in full at the next backup. This is the case even if that modification is only
a single character. This inefficiency reduces the effectiveness of tapes while increasing the quantity of tape
you’ll need to get everything backed up.

Tapes are slow in backing up, slower to restore. Today’s faster tape technologies have improved the
speed of backups; however, that speed continues to be limited by the sheer mechanics of tape management.
An individual file can be spread across multiple tapes, requiring swapping by hand or through a robot device.
Fast-forwarding and rewinding through a tape to locate the correct data further lengthens restore time.

Physical constraints are also a problem. Even worse is the situation where the data you need to restore is on
a tape that’s been rotated offsite. Locating that tape, getting it to the datacenter, and loading it all add time
and inefficiency to the restore process.

Tapes are physical items that can be lost, damaged, stolen, or destroyed. Finally, tapes by nature are
intended to be portable items. They’re designed to be items that can be removed from backup devices and

5
transferred to archive locations. Tape rotation offsite is a common activity, and a best practice in tape backup
architectures. But offsite rotation and even daily handling itself introduces the risk for loss, damage, theft,

6
and destruction.

Today’s datacenters need better solutions that are always connected yet provide offsite replication for
archival purposes.


D2D – The Start of Today’s Backup Solution
Each of the issues noted above are limitations to tapes’ ancient approach to today’s problems. Tapes will
absolutely protect your data, but you’ll be paying a high cost in doing so – both in labor as well as in risk.
That’s why smart IT organizations are looking towards Disk-2-Disk backups as a superior solution. Disk
backups still back up data to an alternate location. The difference is that the alternate solution is an array of
disk drives rather than a stack of tapes. Those disk drives can be located in the same datacenter as your
servers, or elsewhere on the network. A foundation for each of the technology evolutions discussed in the
next sections, disk drives solve many of tape’s problems right off the bat:

Disk backups eliminate the entire concept of backup windows. Tape backups are forced to use the
backup window concept because of their change-oriented nature. A file changes, it needs to be backed
up. Disk backups, on the other hand, operate quite differently. With disks, traditional file-by-file
backups are replaced by a file system filter driver that is installed directly to each backed up machine.
This filter driver intercepts each and every call to a computer’s disk subsystem, noting the change
and instructing the disk backup system to replicate it elsewhere.

Disks are random-read/random-write. Using a filter driver in this manner means that data is constantly
being backed up as it changes. Every update to a file or folder is replicated to the backup device, at the level
of the individual change. These updates are written to disk in a non-linear fashion, which means no waiting
for an empty section of tape. Individual updates are simultaneously entered into a centralized database,
enabling them to be reconstructed at will when restoration is necessary. Capturing data on a per-change
basis instead of a per-file basis significantly reduces the duplication of data that is contained within the
backup set. In short, every piece of data is immediately addressable whenever it is needed.

Today’s disks are enormous, fast, and inexpensive. Disk backups took a long time to overtake tape in
economy due to the initial high cost of large quantities of disk storage. That high cost is no longer a hurdle.


Today’s disks are enormous, fast, and inexpensive. Disk backups took a long time to overtake tape in
economy due to the initial high cost of large quantities of disk storage. That high cost is no longer a hurdle.
With the cost of multiple-hundred gigabyte and even terabyte disk drives dropping to affordability, today’s
disks provide a cost advantage that wasn’t available before.

That affordability also comes with excellent performance as well. Disk-based backup solutions do not place
a heavy reliance on disk performance for overall backup performance. Its low requirements for disk perfor-
mance further highlights how inexpensive disk arrays can be used – or, in some cases, repurposed – for disk
backups at a cost-effective price.

Disks enable emails, files, and entire computers to be restored in the exact same manner. Best of all,
disk-based backups also eliminate the reliance on backing up files and folders. Data can be treated as data.
Recall how disk backups use a filter driver rather than individual file copies to backup data. This action can
be taken a step further, combining it with application awareness. Application-aware backups enable a backup
solution to recognize what kinds of things it is backing up: Emails versus rows in a SQL database, files on a
disk share versus files in a SharePoint implementation, physical machines versus virtual machines.

With application awareness, a disk-based backup solution’s per-change basis recognizes what each disk
transaction means. It can log that change in the correct manner for later retrieval. And, most important,
it means that an individual email, SharePoint item, or SQL database row can be restored just as easily
as an entire computer.

Protecting your
company’s
most critical
data really
means
protecting
your company
itself.

D2D2T – Benefitting from Disks and Preserving
your tape investment
Disks obviously have a set of benefits that go far beyond what you’ve seen with your tape backup investment.
Immediate restorability of data from individual items to entire computers is a fantastic risk reduction. Elimi-
nating duplicated data represents great cost savings. Yet you probably still have a strong investment in tape
backups, one that would be difficult to just throw away.

It is also worth mentioning that tape backups aren’t all bad. They’ve served our industry for most of its early
years, and can continue to provide value for your archival needs.

D2D backup solutions today provide exceptional improvements to backup and dramatically reduce restore
time; however, disk-based backups alone suffer from limitations in their ability to truly archive data. Your
business today in actuality has two different requirements for data preservation:

Requirement #1: Assured and timely restorability. Your business requires the capability to successfully
restore data in a timely fashion. This includes not only file-based data, but also application data such as
individual emails, database records, and even entire computers. Accomplishing this goal is well served
through a disk-based backup solution that can enhance the process as detailed above.

Requirement #2: Long-term data archival. Whether your business falls under the purview of regulatory
compliance or not, you need the ability to store copies long term. Those long term copies can be for historical
purposes, or to simply preserve the actions and activities of your company over time. While entirely possible
to store via disk-based backups, these long term copies are perfectly suited for tape-based backups.

The combination of these two needs highlights the importance of considering multiple approaches in your
overall data preservation policies. Disk-based backup solutions are perfectly suited to be inserted between
your servers and your existing tape infrastructure. From that position, a disk-based backup solution can
handle the day-to-day operational needs for data preservation. Following along behind is the tape-based
backup solution, essentially “backing up the backup solution” for archival and long-term storage purposes.

There is a better way…and that better way is called hybrid backups.


D2D2C – A Solution for Eliminating
Tapes Entirely
Backup technologies are indeed evolving at a dramatic pace, one that matches today’s cloud-based
infrastructures. It is exactly within that cloud infrastructure where the most revolutionary of backup
and archival solutions are finding a home. Disk-2-Disk-2-Cloud has quickly matured to become a viable
option for many organizations who wish to eliminate the administrative cost and risk of tapes entirely.

In a D2D2C architecture, locally-homed disk-based backups are used to handle the daily operations of system
backups. These disk-based backups are no different than those discussed above, and enjoy all of the benefits
gained through the move away from tape.

Different here is in how the backups are themselves backed up. Unlike D2D2T, which preserves a company’s
tape infrastructure, D2D2C instead offloads the archival responsibility to a cloud service on the Internet. That
cloud service, hosted by a top-tier partner and enjoying levels of security above and beyond traditional offsite
rotation, immediately becomes a trusted location for the secondary storage of data. Protected, encrypted,
authenticated, and available anywhere with an Internet connection and the correct security controls, your
data now becomes pervasively available. You also gain a set of benefits not easily attainable inside your
own datacenter:

Entire site data protection. In the worst of catastrophes, companies have lost entire datacenters. Whether
due to naturally-occurring events or outside influences, protecting your data means keeping it safely archived
in secondary locations. Cloud-based solutions provide that secondary location. At the same time, they also
provide cloud-based restore capability. In the case of an entire-site loss, your data as well as the servers
that host it can be quickly transferred over the Internet to a secondary location. Leveraging strong and
multi-factor authentication as well as encryption in transit, data traversing the wire remains secure on its way
to archival. Protective measures at the cloud services location further secure that data through data isolation,
activity monitoring, and further layers of redundancy.

Conversion of CAPEX to OPEX. A tape investment is also a capital investment, one that occasionally incurs
unexpected costs. Those costs can be challenging for companies with restricted budgets. A better solution can
be the conversion of tape’s capital expenditure into a cloud services’ recurring and predictable operational
expenditure. Cloud services eliminate the risk of unexpected expenses while providing a recognizable and
predictable cost that fits well into budgeting guidelines.

A scope of infrastructure you can’t build yourself. Cloud services are financially feasible in part due
to economies of scale. By aggregating the backup and restore needs of multiple customers into a highly-
engineered and specifically-designed solution, cloud services can provide a level of infrastructure you simply
can’t build in your own datacenter. Many cloud services provide advanced protection measures such as
underground data storage, multiple levels of access control, multi-factor and strong authentication, and
activity monitoring that ensures only the right set of eyes see your sensitive data.


The Hybrid Approach – Consolidating
Disk, Tape, and Cloud
This entire introduction paves the way for backup’s greatest evolution, one that integrates the benefits of D2D
with both tape and cloud backups. This everything-to-everyone model is called the Hybrid Approach.

The Hybrid Approach uses today’s technologies to gain the best of all worlds. Some data simply doesn’t
make sense for Cloud backup. Perhaps that data is limited by regulation, by security policy, or falls under
special data ownership requirements. Some data sets are too large or change too rapidly to be efficiently
backed up to the Cloud. In these cases, Cloud services might not make the best sense for certain sets of
data. In those cases, retaining your existing tape backup infrastructure can be an excellent extension of
its operational lifecycle.

Other data fits perfectly in the cloud backup approach, providing the highest levels of protection in
combination with worldwide restorability. The hybrid approach right-sizes the best-suited technology
to your data retention needs. Leveraging an integrated toolset that enables tagging data to the correct
locations, software that uses the Hybrid Approach creates a single location where all data is backed up,
cataloged, access controlled, and eventually restored.


Iron Mountain CloudRecovery is the Hybrid Approach
Integrated with Microsoft’s System Center Data Protection Manager (DPM) 2007 SP1 and 2010 releases,
Iron Mountain CloudRecovery is a one-stop solution for implementing the Hybrid Approach in your datacenter
today. Its services provides full-data protection for files and folders, as well as Microsoft applications such as
Hyper-V, SharePoint, Exchange, and SQL Server.

Iron Mountain CloudRecovery extends DPM 2010 by providing an offsite backup location that is
buried deeply underground and protected through multiple and layered security mechanisms.
Microsoft DPM 2010 provides the workflow framework in which data is tagged for primary disk
backup as well as secondary backup to either the Cloud or to a local tape backup device. Using
Iron Mountain CloudRecovery and Microsoft DPM 2010, any application, file share, or entire
computer can be backed up and archived offsite for as few as thirty days or as many as seven
years. Through DPM 2010’s central console, data in primary storage or in onsite or offsite
secondary storage can be seamlessly restored to any server at any time.

For more information, check out: http://www.ironmountain.com/cloudrecovery
Copyright © 2010 Concentrated Technology. All rights reserved.

Greg Shields is an independent author, speaker, and IT
consultant, as well as a Partner and Principal Technologist
with Concentrated Technology (www.ConcentratedTech.
com). With 15 years in information technology, Greg has
developed extensive experience in systems administration,
engineering, and architecture specializing in Microsoft OS,
remote application, systems management, and virtualization
technologies. He is a Contributing Editor and columnist for
TechNet Magazine and Redmond Magazine, and serves as
the Series Editor for Realtime Publishers, the world’s lead-
ing provider of high-quality content for the IT market. Greg
is a highly sought-after and top-ranked speaker for both live
and recorded events, and is seen regularly at conferences like
TechMentor Events, Microsoft Tech Ed, VMworld, and more. He
is a multiple recipient of Microsoft “Most Valuable Professional”
award as well as VMware’s vExpert award.


2010 data protection best practices

More Related Content

What's hot (18)

Viewers also liked (17)

Similar to 2010 data protection best practices (20)

2010 data protection best practices