When it comes to business continuity, if you are proactive and foresightful enough to set up a plan in advance of disaster, it is of course tempting to set up a system that is absolutely bulletproof. Unfortunately, it often transpires that there are tradeoffs to make in order to square your desired backup plan with the limitation of your budget.
Active or Passive? What’s the Difference?
There are two main structures for backup architecture: active-active and active-passive. Active-active architecture is optimized for uptime. In this setup, off-site servers are synched with the onsite server, allowing for near-instantaneous replication of data. Usually, this is set up so that when one or more servers fail, the load is automatically rebalanced to the rest of the servers. Individual applications can also be set up with the same strategy, as not every application will cause substantial issues if there is downtime.
Active-passive architecture is less complex. In this setup, a separate set of critical infrastructure is maintained off-site, but it sits idle until such a time as it is needed. This setup does have some downtime associated with it, and may also allow some data loss. But it also has its benefits, as we will later show.
What Are You Prepared To Lose?
There’s no right answer when it comes to designing a disaster recovery strategy. But there are a couple of metrics that organizations can use to decide the best course of action.
- Recovery Point Objective (RPO)
- Recovery Time Objective (RTO)
RPO is the affordable amount of data an organization is prepared to lose in the event of an outage. This number essentially results in the minimum time interval for data replication. So if an organization can soldier losing 4 hours worth of data, then it can run backups every 4 hours. If even a few minutes of lost data is intolerable, then the replication will run every 15 minutes, or even continuously if necessary.
RTO is the affordable amount of downtime permissible between an outage and the return to functionality. This metric will suggest the type of equipment backups, their class, and the speed of your connection to those backups.
Naturally, the smaller you need those numbers to be, the more that it will cost. If you go full active-active, maintaining far more equipment and capacity than you need at all times in order to be fully prepared for an outage, you’re not just paying for that availability and that hardware, you’re also paying for duplicate licensing for all those machines as well. The numbers add up.
That’s what we mean about the perfect backup meeting the hard limitations of the budget.
Deployment Options
Luckily, there is a little wiggle room for every setup. Let’s go over a few options for disaster recovery deployment systems.
- Synchronous Replication using active-active arrays—this is the Cadillac of systems, but it’s going to cost you. It’s complicated to implement as well.
- Asynchronous Replication using active-passive arrays—this setup takes backup snapshots at intervals but otherwise remains idle. This is a relatively simple setup, with the RTO being about as long as it takes the system to boot after failover, and the RPO being as far back as the last snapshot. Cost-effective for most.
- Combination of the options depending on seasonal needs and-or critical applications.
There are more low-cost solutions, such as file backups, but these typically do not suffice for RTO.
If you’re considering setting up or upgrading your current BCDR strategy, we can help you triage your needs against your budget and help you set up all the failover and recovery options with our best-in-class Business Continuity service. We walk you through, step by step, from strategy and pricing to implementation and testing. Simply give us a call, or write to sales@atlanticmetro.net today! Business continuity shouldn’t be left to chance.
The Technical Challenges of Big Data