- Do You Want Chaos With That?
Every day, we make choices. Some choices you make without thinking. Some choices are made for you. Some choices cost you sleepless nights, friends and opportunities for eternal happiness. So, when somebody asks, “Do you want fries with that?” consider the long-term implications.
In backup, choices matter more than ever. Over the next few years, we’ll witness the complete transformation of the backup industry, and you don’t want to find yourself with the data protection equivalent of a Zune MP3 player, HP Netbook, or a Breece Hill Q2.15 tape stacker.
Not surprisingly, at the epicenter of the change in the backup industry is VMware. As it expands into more critical applications, it is disrupting existing backup and storage processes, and how you respond (i.e., the decision you make) will set the tone for the rest of your environment for years to come. The right choices will make for a smooth transformation process, but the wrong choices could put you on course for more silos and chaos. (And all that pressure comes before I ask whether you want fries!)
Ask Yourself Three Questions…
When considering your VM protection, ask yourself three questions. Ask yourself these same questions when you’re revisiting your application and NAS protection approaches, trying to figure out how best to leverage the cloud and even when you’re deciding what to have for lunch:
- How am I leveraging the intelligence in the data source and data destination?
Ten years ago, most of the backup intelligence resided in the backup software. Application servers did not optimize for data protection and nobody ever accused tape of being intelligent. The growth in data capacity, the shift to 24×7 operations and the reduction in spare resources has changed the two endpoints for data protection. Applications and virtual servers have begun to optimize for data protection. For example, VMware has Changed Block Tracking (CBT) and Oracle supports Incremental Merge and Block Change Tracking. Meanwhile, backup appliances are replacing tape. These intelligent arrays support deduplication, replication and virtual synthetic backups. By leveraging the optimizations on both sides of the data flow, you can reduce VM backup and recovery times from days to minutes. The first step to this dramatic change is simple – you have to find the technology that leverages the intelligence at the data endpoints.
- Will I deploy a point product for protecting this data, or make it part of a central architecture?
Virtually every customer I meet uses multiple backup mechanisms, and they struggle to gain control of their environments – the location, retention, security and cost of their backups. Why are there so many backup silos? Engineers love to solve technical problems; it’s why we all joined IT. Therefore, if the backup team cannot meet the needs of their consumers (e.g., VM admins, application admins, etc.), they are happy to solve the problem themselves. At EMCWorld in May, I talked about Ecolab. I explained how Ecolab’s VM administrators were so frustrated with their backup service levels that they explored options that ranged from building their own tool to purchasing a new solution. They involved the backup team only after they made their choice for EMC Avamar; after all, somebody needed to actually run the backups! Fortunately, Ecolab found that EMC’s backup solutions can optimize their entire complex, distributed environment. Not all customers have been so lucky. The backup team must accept that dissatisfied VM administrators will search for a better answer. Therefore, they need an architecture that can leverage both the optimizations (e.g. CBT) and interfaces that their VM and application administrators will want to use (e.g., vCenter and Oracle RMAN).
- What recovery workflows will I need to support?
Virtually every customer I meet uses multiple recovery solutions for their VMs. They use their backup product to run operational and long-term recoveries. They deploy snapshots to roll back VMs after . They depend on storage replication for offsite disaster recovery. In addition to the complexity and cost of operations, each approach comes with significant limitations. Backup products lock up data into proprietary formats. Snapshots recover storage containers that contain dozens or hundreds of VMs, not individual VMs. Storage replication is bandwidth and cost heavy. We believe customers need to ask for more. One solution should support all recovery workflows – operational, long-term, roll-back and offsite DR. That solution should keep data in open formats, enable VM-granular protection and reduce the cost of end-to-end disaster recovery. Customers should accept nothing less.
These questions are simple, but they will help you make the right choice for protection – not just for VMs but for your whole environment. These questions also apply whether you’re a global Fortune 100 organization, a distributed enterprise, or a fast growing company. In any of these instances, you will face the challenge of VM backups – as a company, a workgroup, or a site. In all of these instances, you want the technology intelligence, interface, and recovery flexibility that meet your needs today (e.g. VMware VDP) that can scale to the future (e.g. EMC Avamar).
More than ever, your choices matter. Still, there is something more important than the choice you make – how you respond to the choices you (or somebody else in your organization) made. You’ve already installed some solution(s) for VMware backup. Whether you’ve made a choice that you’re happy with or one you regret – you still have the opportunity to make a new choice and order those fries.
- CTO Connection
Thoughts from Stephen Manley
The Right Architecture Is Priceless, Part I
Evolving data protection technology and expanding requirements have completely transformed the backup industry. Unfortunately, with such rapid change, many organizations have fallen into the chaos of an accidental architecture.