![\"Pug](\"https:\/\/nsrd.info\/blog\/wp-content\/uploads\/2015\/09\/bigStock-DPUG.jpg\")
<\/a><\/p>\nIf you’re interested in joining and participating and based in Melbourne, you can find details for the user group over at Meetup<\/a>.<\/p>\n Our first presentation was on\u00a0Wednesday 9 September, and EMC Melbourne were kind enough to provide the office space for the session. That being said, DPUG is not about EMC products – it’s designed to be\u00a0a vendor neutral community forum to discuss techniques,\u00a0strategies and best practices relating to data\u00a0protection.<\/p>\n Starting DPUG was a healthy reminder that\u00a0data protection<\/em> is an overloaded term in the IT\u00a0industry. To those of us who work within\u00a0data storage and more broadly, IT\u00a0infrastructure,\u00a0data protection<\/em>\u00a0covers concepts such as\u00a0backup and recovery,\u00a0continuous availability, continuous data protection, replication, snapshots and so on. For people who work at the application layer\u00a0or communication\u00a0layer though, data\u00a0protection is almost invariably interpreted to be something like\u00a0security,\u00a0data privacy<\/em>\u00a0or intrusion detection\/threat mitigation<\/em>.\u00a0Data protection is a term we share with other areas of the industry. In the end it’s\u00a0all<\/em> data protection, but it has two very different areas of focus.<\/p>\n Our first session was about VMware Data Protection.\u00a0We’re now seeing a\u00a0very high percentage of\u00a0virtualisation within most businesses \u2013 it’s not uncommon to see\u00a080% or 90% virtualisation now, and many companies are continuing to pursue a strategy of achieving 100% system and infrastructure virtualisation.<\/p>\n In the VMware Data Protection presentation I walked\u00a0the audience through a history of\u00a0how the industry overall has\u00a0protected virtual machines since their inception in the midrange space. First, we started with treating virtual machines like regular hosts \u2013 installing agents on each virtual machine and backing it up as if it were no different from\u00a0a physical host.\u00a0That provides a high degree of granularity and\u00a0flexibility, but as we know, virtualisation is about cooperative resource sharing, whereas traditional backups are about minimising the time it takes to get data from\u00a0the client into the protection storage.\u00a0There’s not a lot of compatibility between “cooperative resource sharing” and “minimising the time it takes to get data from the client…”, and a poorly designed backup strategy using in-guest backup agents can bring\u00a0virtual infrastructure to a screaming halt \u2013 even today.<\/p>\n The next attempt to provide a comprehensive solution for backing up\u00a0virtual machines saw businesses installing\u00a0backup agent software on the hypervisors, and writing custom scripts to snapshot virtual machines prior to copying them to protection storage. This was usually error prone and when you stop to think about how virtual machines are usually just very big files, it meant that a single change within a virtual machine would trigger a new full backup every time. Once technology such as VMotion became available these techniques became difficult if not impossible to maintain \u2013 you could not really predict\u00a0where<\/em> a virtual machine would be for backups at any given time. What’s more, hypervisors are a bit like NAS appliances \u2013 they’re designed to do one thing really well, and you shouldn’t be trying to install third party software on them.<\/p>\n The solution was an API based approach, of course.\u00a0While different in practice, you can equate the API approach of VMware backups to the NDMP approach of NAS. The virtualisation system provides an integration point for backup software to use, and leveraging that, backup\u00a0products are able to streamline the\u00a0data protection process with image level backups\u00a0and<\/em> file level recoveries from those image level backups.<\/p>\n This is something that\u00a0NetWorker for instance has been doing for some time \u2013 most\u00a0recently with VBA.\u00a0VBA is something I’ve covered a few\u00a0times over\u00a0the last twelve months (Current state of Virtual Machine Backups in NetWorker<\/a>, NetWorker 8.2 and VBA Instant-Access<\/a>, and Testing and Debugging an Emergency Restore<\/a>, for instance).<\/p>\n VMware offers its own version of VBA as well so that businesses (particularly smaller ones)\u00a0can still protect their environments. It used\u00a0to be split into VDP and VDP\/A, but as of vSphere 6 Essentials,\u00a0those\u00a0options have been combined into a single (free) VDP.\u00a0VDP can’t do everything VBA can do\u00a0\u2013 for example,\u00a0VDP can’t:<\/p>\n As a means of demonstrating\u00a0some of the advantages of virtual machine image level backups though, VDP is\u00a0useful, and that’s what I used in the DPUG session earlier this month. And now, after taking the plunge and investing in some screen recording software, I’ve made three of the demos from the DPUG session available for viewing. If you’re using VBA already\u00a0you’ll be familiar with all of these. However, if you’ve not yet taken the plunge in utilising VBA for your backup environment, check them out \u2013 while the demos show the VMware Data\u00a0Protection Appliance (VDP) in use, they’re equally applicable and in fact it’s the same process for a VBA install in each\u00a0situation.<\/p>\n Creating and executing a protection policy:\n
\n