The devastation wrought by superstorm Sandy is a stark reminder of just how fragile our environment is versus the power of nature. In just a couple of days, a single storm disrupted the lives of so many people and paralyzed a large number of businesses even days after it passed. In the wake of the storm, one of the questions on my mind is, how can we help businesses remain functional after such a massive hit regardless of their size?
Working for an Application Delivery Controller company, I’m no stranger to disaster recovery. Most often, disaster recovery is initiated by large enterprises that invest big money to build entire backup datacenters with the ability to automatically provide all online services in case their main datacenters become unavailable. A reality faced by many businesses in and around New York after the storm.
But what about smaller businesses that can’t afford to build and continuously maintain secondary datacenters in remote locations? It’s clear that smaller businesses need a more cost effective solution that can provide the same basic function – continuous operation of all internal and customer-facing mission critical (and revenue generating) applications in case their main (and only) datacenter becomes completely unavailable.
Current technology offers one possible solution. Cloud services, such as infrastructure as a service (IaaS), allow organizations to buy computing resources, receive Internet connectivity and host mission critical applications in a private virtual datacenter. However, while IaaS is the basic building block, there are still two additional components required to complete the disaster recovery (DR) solution – data replication to the redundant site and traffic steering to the DR site in case of disaster.
Data replication solutions:
Data replication is an essential component of the overall disaster recovery solution. Fortunately, there are various off the shelf software packages that help with data replication. From real-time data backup that provides continuous data protection (CDP) to advanced data replication between main and DR datacenter sites, there are plenty of ways for small businesses to ensure the consistency and availability of their data even in the event of a disaster.
Automatic traffic steering across sites:
The solution for automatic traffic steering across datacenters has been around for quite some time through load balancing/ADC devices. Using Global Server Load Balancing technology (GSLB), ADCs can monitor the local health and availability of the application server, as well as receive information from remote ADCs regarding the health and availability of applications in remote sites. Based on this combined information, traffic can be easily rerouted to the most available application server. In case of total datacenter failure, all traffic will be automatically rerouted to the DR datacenter – providing continuous operation of all of an enterprise’s mission critical applications.
To enable GSLB across datacenters, organizations have to install the same ADC type at both (or more) datacenters. However, most IaaS cloud providers won’t allow their customers to install their own hardware onsite since it contradicts the idea of cloud computing!
One possible solution is in the form of a soft ADC running as a virtual appliance on a general purpose computing resource. Just like any other software, a soft ADC virtual machine can be installed in the cloud, providing all the services required for a DR solution. But for such a solution to work, your ADC vendor has to meet the following requirements:
- Provide both hardware based and software based ADC form factors
- All ADC form factors must have identical functionality
- GSLB must be supported across form factors (i.e. physical and soft ADCs)
The good news is that this cloud based DR solution is already available with Radware’s Alteon ADC solution, which is considered a Magic Quadrant Leader. Enterprises of any size, using any of Radware’s Alteon appliances can cost effectively build a cloud based DR solution and ensure continuous operation of their mission critical applications, even in the most severe cases where their main datacenter becomes completely unavailable.