Modernizing Backup and Recovery in a Virtual World

I thought following article from early last year summed up the why’s best. Reprinted here in its complete form.

Virtual Strategy Magazine: “Modernizing Backup & Recovery in a Virtual World”
January 28, 2011
By Rob Emsley

Companies easily get started with server virtualization – typically internal IT applications. Having tasted the obvious benefits, they quickly embark upon a broader project – virtualizing as many servers and business applications as possible. And then they hit a wall; virtual server sprawl makes their infrastructure hard to manage; bottlenecks spring up due to resource contention; and backup windows can no longer be met. The solution is innovation – better management and more robust backup and recovery processes that are designed specifically for today’s IT landscape.

“Data protection paradigms and architectures (especially backup and recovery) must change in order to adapt to virtualized environments.” IDC’s Worldwide Storage for the Virtualized Environment Taxonomy, 2009, December 2009.

Navigating the Three Phases of Virtualization

There are three primary phases of virtualization. The first stage is known as IT Production. The IT department deploys virtualization for internal applications such as web servers and caching appliances. While important, these areas could not be classified as mission critical. The aim for IT is to use virtualization to enhance cost efficiency. In this phase, virtualization rates range anywhere from 0% to 30%.

Phase 2 can be characterized as Business Production. At this stage, the goal is to align virtualization of IT resources to business goals. Instead of just making IT more efficient, the intention here is to extend those efficiencies into the organization as a whole by streamlining business processes. By this point, virtualization ranges from 30% to 70% and companies realize that a traditional approach to backup and recovery is no longer enough.

Phase 3, known as IT as a Service (ITaaS), is where true business agility is unleashed. The organization achieves the greatest possible gains from virtualization by reason of optimized processes, low transactional latencies and IT cost efficiency. However, this can only occur once organizations have passed the 70% virtualization mark and have adopted a next-generation approach to backup and recovery.

Hitting the Wall with Traditional Backup Processes

Organizations in the midst of the second phase sooner or later arrive at the same conclusion: greater levels of virtualization are accompanied with mushrooming data protection needs. Why? Under the old physical world paradigm, backup was done on a series of servers, each of which was relatively poorly utilized in terms of CPU, RAM and disk space. Consequently, abundant resources existed with which to accomplish timely backups. CPU’s could spend most of the night solely on backup processes. As long as organizational applications were ready at the start of the next day, the business was unaffected.

But as the virtualization ratio rises, last-generation backup applications eventually reach a breaking point. The reason is simple. Consolidation ratios of 20 virtual machines (VMs) onto one physical server are commonplace. That means 20 times the backup overhead on that one server, as well as any additional overhead consumed by VMware. On top of CPU/memory resources, bandwidth is also far more constrained in an environment containing many VMs. Attempting to utilize agents running on each VM, for example, becomes untenable due to extreme resource contention.

At the same time, virtualization facilitates data growth. According to Enterprise Strategy Group, one of the byproducts of server virtualization is growth in storage volume due to proliferation and redundancy. This is particularly the case in organizations with high percentages of virtual servers and large volumes of data to protect. After all, the efficient way to safeguard VMs is to back up the entire file – the operating system, applications and data. This approach enables better mobility, removes hardware dependency and enables rapid recovery. But as it is so easy to backup complete images, the downside is that data storage demands increase markedly.

Thus there is a need to transition to a proactive approach to VM backup. In fact, it is a necessity if an organization hopes to move through the stages of virtualization and attain business agility. Those relying on outdated backup processes will inevitably bog down in Phase 2 and fail to progress to Phase 3.

Backup Wish List for Virtualization Success

Success in today’s highly virtualized data centers, then, has to resolve several critical areas.

A. Address resource contention

B. Contain storage growth

C. Enable more effective DR

D. Provide business value via lower backup infrastructure costs, improved operations and reduced risk

E. Integrate with VMware components such as the VMware vStorage API and vCenter Server

Those that plod onward hoping that a business-as-usual approach to backup will suffice are likely to descend into a form of virtual purgatory – no matter how hard they work and how much hardware they throw at the problem, their backups will continue to either be too slow or too expensive.

Further, reliability becomes even more crucial as more and more VMs are hosted on physical servers – the consequences of a failure increase dramatically. Therefore, virtualized operations demand far more effective DR as more VMs and more applications are potentially impacted.

The challenge is to simplify backups, while reducing risk and lowering costs. This can be achieved through a combination of deduplication and a redesign of backup and recovery that is tailored to the virtual world.

Deduplication and Virtualization-Ready Backup

Deduplication can be accomplished in a couple of different ways. Deduplication storage systems are particularly effective at containing explosive storage growth and providing more reliable DR via replication. By substituting disk for tape and working with existing backup software, they offer significant gains in the enterprise. That’s why deduplication appliances have experienced one of the most rapid adoption curves in storage history.

What they don’t address, though, is resource contention. Deduplication at the source goes a long way towards solving resource contention challenges. Instead of transmitting all data over the network, only unique data is allowed to consume bandwidth. This frees up a tremendous amount of CPU and memory resources to ensure backup windows remain short.

But deduplication is only part of the puzzle. It must be accompanied by the latest backup software applications that are designed to work within a VMware environment. Instead of backing up in the same way as in the physical world –– next-generation backup software integrates fully with VMware to provide faster backup and ease of management while minimizing resource usage.

VMware vSphere 4, for example, introduced several features that greatly improve data protection processes by fully leveraging features of the hypervisor.

A. The vStorage API for Data Protection (VADP) has replaced VMware Consolidated Backup (VCB). VADP allows a live system image snapshot to be captured without impacting applications or over-taxing the host’s CPU. Further, a physical proxy server is no longer required. Virtual proxies can be used.

B. Changed block tracking (CBT) tracks changed blocks of a virtual machine’s virtual disk. This helps backup software to identify and copy only those blocks changed since the last backup.

The combination of deduplication in tandem with VMware-specific backup provides the ideal working environment for virtualized platforms.


Only those organizations which redesign their backup and recovery systems using virtualization-ready technology will be able to avoid the inevitable bottlenecks and frustrations that occur in Phase 2 of virtualization. By adopting a full range of approaches to deduplication along with retooled backup software that takes full advantage of vSphere 4 features, organizations can at last graduate up to Phase 3. They will be able to combine the best features of traditional and next-generation back up technology to fully protect their environments.

That translates directly into increased consolidation ratios, 90% faster backups, 40 to 60x data reduction and a 90% reduction in replication bandwidth. In addition, it extends business value by lowering backup infrastructure costs, simplifying management (by enabling backup to be monitored through vSphere), improved operations and reduced overall risk.

It makes sense, therefore, for organization embarking upon the virtualization adventure to take a cold hard look at their data protection processes. By proactively addressing the needs of backup and recovery in a virtual environment, enterprises can not only avoid the burden of bottlenecked processes, they can lay the foundation for a smooth transition into the third phase of virtualization – where they gain the agility needed to compete in today’s challenging times.
Rob Emsley – Senior Director of Product Marketing at EMC
Rob Emsley is senior director of product marketing at EMC Corp. Rob is responsible for marketing the EMC Backup Software portfolio. His organization drives activities to bring products to market that help customer’s gain value through backup redesign. Rob joined EMC in 2004 and has managed product marketing for several of the software acquisitions that help deliver information infrastructure solutions.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.