Just as important as planning for scale is being prepared for when everything begins, proceeds, and completely falls apart. Although the causes of infrastructure disruptions may vary, whether caused by heavy load on the servers, natural disasters, or unexpected technical issues, having a plan to mitigate the difficulties is imperative. Should that plan not be enough, a full-scale blueprint for the scenario after the worst-case scenario is in order. While what constitutes a disaster for each company may vary—for a mobile app, the company’s site going down may not be as devastating as it may be for a site subscription-based page, having a business continuity plan for challenging times and a disaster recovery plan for catastrophic times is not only important, but as one can always tell while in the midst of it, imperative.
When beginning to draft a disaster recovery plan, two important factors to consider are the length of the recovery time and the cost of maintaining the appropriate safeguards. The availability of cloud computing has reduced both.
Many cloud providers offer free or low cost backups. The presence of these alone should provide some security in case of a loss of data. However, depending on the importance of your site and content, the minimal option, of site backups, would not be enough in the case of a digital disaster.
Although consistent backups, whether done by the cloud hosting provider or manually by the user, provide some security, the low price of cloud computing opens up some other avenues for creating fault-tolerant sites. Cold sites, replicas of the current web page that are not currently online, are the cheapest route to getting back online in the case of a disaster. As cloud hosting offers a very fast spin up time, what used to be the slowest of all of the recovery methods can now be accomplished in hours rather than days. Additionally, most cloud hosting providers allow hourly usage of their servers—this further reduces the expense of the cold site, since the charges are only incurred if the site is up and running.
At the same time, depending on the required speed at which the digital disaster recovery needs to be completed, a hot site may be a better alternative. This can be set up with a cloud-hosting provider across several of their data centers and locations, or, to be even more secure, between several providers.
The lower cost of cloud computing itself (versus traditional dedicated servers) as well as the flexibility in datacenter locations and the fast spin up time has made digital disaster recovery much more affordable and brisk.
However, when looking into the price of cloud hosting, companies should avoid falling into the pitfall of trying to take the savings and ‘run,’ so to speak. Especially with economy cloud hosts, if a site is worth more than the low cost of hosting it each month, the difference should be invested into some sort of digital disaster recovery preparation.
The common thread throughout this piece is the responsibility of developing a business continuity plan and disaster recovery plan lies squarely on the shoulders of the developers setting up the site or app. With a couple of exceptions, the cloud is mostly unmanaged. This passes both savings and additional work onto the user. Customers should never expect that the their hosting company can simply restore their data, keep their website running, or even not be washed away in the case of a natural disaster.
Cloud computing as an industry has eased some of the difficulties of digital disaster recovery, but only when a startup takes the time to create an effective recovery plan that works for them, can they be assured that their data, whether it’ll take minutes or hours, will eventually be back online.
Mitch Wainer is the co-founder of DigitalOcean, a New York-based cloud server and hosting provider. DigitalOcean is headquartered in New York City and graduated from the TechStars startup accelerator program. Featured in Forbes, Pando Daily, TechCrunch, Venture Beat, and The Next Web, the company has launched over 190,000 cloud servers since inception.