Superlative SLA: Why ‘Good Enough’ Isn’t Enough in a Disaster
Companies can’t afford to underestimate the impact of network failure. As noted by GCN, enterprises experience between 10 and 20 hours of unplanned downtime each year, even in the absence of natural disasters. While business continuity and recovery plans have become standard operating procedure for many organizations, it’s worth taking a hard look at hidden costs, technology considerations and the need for a service-level agreement (SLA) more superlative than standard.
The Price Is Right?
When enterprises think about cloud backup, several common data points emerge: storage, speed and support. If companies are willing to pay for all three, shouldn’t the result be straightforward disaster recovery? According to IT Business Edge, however, it’s easy to miss critical costs such as the pricey nature of cloud storage. Why? Because the storage space needed for a “grandfather-father-son” backup solution using daily backups and first-in-first-out access is much greater than the storage required for on-premises and in-use data. The result is costs that are far greater than anticipated.
Speed also comes with a greater price tag than many companies realize. While providers may be committed to delivering backup data as quickly as possible, repopulating entire servers with terabytes of data can take days or weeks, even over a solid connection. One way to solve this problem is with reverse seeding, which sees physical media overnighted and installed to jump-start the process of data recovery. Not surprisingly, however, this service doesn’t come standard.
When it comes to support, expect to pay a premium for local, expert IT talent available 24/7. While this may be available on-demand rather than as a long-term service, be prepared for substantially increased rates if a disaster occurs over the holidays or affects multiple enterprises over a wide geographic area.
Governing any disaster recovery operation or cloud initiative is the SLA. It spells out the responsibilities of both provider and client in addition to specifying elements such as mandated recovery time after a disaster occurs, costs (if any) to move data from the current cloud provider to another and specific storage and retrieval conditions of data if a business continuity plan is activated.
The problem? According to TechTarget, because cloud computing SLAs haven’t evolved into an industry standard, there’s a lack of specific terminology or common clauses that form the basis of any good agreement. As a result, many companies accept the cookie-cutter SLA provided by a prospective vendor — a document that invariably favors the vendor and leaves the client with little recourse.
It’s therefore critical to bring in both legal and tech experts to review a service-level agreement and ensure specific clauses about storage, data transfer and recovery costs are included. In addition, it’s important to spell out penalties for failure to meet disaster recovery objectives in addition to drawing a clear line between infrastructure owned by the provider and apps built by the client.
Make no mistake — disasters can happen. SearchDataCenter reported that in early January 2016, a popular airline experienced major delays when its cloud provider experienced a three-hour power outage. At no time did the airline’s network infrastructure failover to any type of disaster recovery server, leaving customers frustrated and tech experts wondering what happened to the company’s SLA.
The takeaway? Reliable disaster recovery costs more than expected. But even big spending doesn’t guarantee recovery success unless enterprises weigh their options, review their terms and design SLAs as the foundation of success.