We talk to Brennan IT Senior Projects Engineer, George Parker, about the technologies that make Disaster Recovery (DR)-friendly environments.
With everything from natural crises to ransom ware attacks making news recently, it would be good to think that every organisation has revisited their DR plan. According to George Parker, though, lack of resources and complexity can mean there are gaps.
‘A number of businesses have made a plan, but haven’t fully tested it, which means they don’t really know if it will work as they hope.’
Complexity is something that must be tackled urgently if DR plans are to stand up to testing.
‘If you don’t have visibility of everything, it is very hard to make sure it is backed up, and you can’t be sure it is protected in your DR plan,’ said George.
Solutions like hybrid IT are easing the pain, and making strong DR plans more feasible, according to George. The advent of mature cloud-computing solution providers means businesses can make better technical decisions on where to best place their infrastructure, while technologies such as Hyper-converged gives visibility of the entire environment from a single point.
‘When you can see it, you can plan for it,’ said George.
This means, he said, that to get a DR plan right, you first must take a step back.
First Thing’s First
‘First, you need to decide where is the best place to run your applications. For any application or workload that contains sensitive data, such as your customer’s personal details or credit card information, it is best to build your own private cloud, either on-premise or with your managed service provider,’ said George.
This private cloud might also accommodate intellectual property, and data that is required by legislation to reside locally. It also makes more sense for applications that demand very low latency or require high-bandwidth consumption.
A hyper-converged solution using current technology would incorporate highly-dense,10GB-plus uplinks from the servers divide and share traffic produced by the virtual machines (VMs), meaning no more spaghetti-style cabling to test the sanity of the IT department. Faced with putting a DR plan into action, nobody wants to be untangling a web of wires.
Flash-first storage systems are no longer only for those with huge budgets, signalling a departure from the usual bottlenecks for organisations with more modest means. They may use externally connected SAS or fibre-channel fabric storage arrays. Alternatively, software defined storage is increasingly an option, using software to build commodity servers into a highly-available converged storage cluster – complete with features found in traditional external storage.
Blade systems can, said George, be an attractive way of building this infrastructure.
‘The network and storage traffic can be virtualised and abstracted by using high-speed adaptors within each blade, interconnected to a common back-plane using “Flex-fabric” switches that segregate storage and network traffic.’
Tools of the Trade
To reap the benefits, George recommends using proven, industry-leading hypervisor solutions –HPE’s SimpliVity already has a steady footprint in Australia’s mid-sized businesses.
Tools like Office productivity, email, intranets, and archived data are best placed in your chosen cloud environment, said George. That takes away yet more complexity in order to focus on keeping data safe.
‘Your backup solution can back up data directly from storage snapshots, even backups of entire VMs with granular restore ability where you need to keep multiple backup copies.’
Your primary backup copy is typically held on-premise near source production VMs for short-term retention and high-speed restore times. Secondary copies are sent seamlessly to your offsite cloud backup repository.
‘This type of environment enables you to recover from ransom ware attacks. Often, once viruses have spread within your private cloud, they also infect your backup server and local backup repositories, so having secondary backup copies elsewhere, in multiple locations, means you can recover your systems to the point in time before the attack occurred,’ said George.
To Host, or Not to Host?
Disaster Recovery is greatly simplified in a virtualised cloud world, said George. By using “host-based” or “Hypervisor-based” DR solutions, you are free to use a mixture of server and storage hardware in your Production and DR facilities. Virtual machines can then be replicated to your own DR facility, to a cloud-providers facility or a combination of both. You may even, said George, opt for DR as a service (DRaaS), and let someone else take care of the rest.
Ultimately, although Brennan IT provides many solutions to protect your systems, George said it is good practice that will keep your systems safe. You need to ‘test, regularly test, all of your systems’ thoroughly to support your DR plan, preferably with a fresh set of eyes – some cloud providers or technology partners offer this service. And you need to train staff regularly.
‘Social engineering is powerful,’ said George.
‘Cyber-crime can be very convincing, so you need to make sure your people know what to look for.’
Need to know more about making your environment DR-friendly? Contact the experienced Brennan IT team.