Cyber-attacks and natural disasters like hurricanes, floods, and fires always grab the headlines. But, what causes the high majority of business interruptions are ordinary, everyday events such as power failures, human error, and faulty software or hardware. As a result, most companies will experience a significant IT disruption at some point. Many are turning to cloud-based backups as a way of providing faster, and more reliable recovery. Often those signing up for cloud infrastructure assume the snapshots or backups offered by their Managed-Service-Provider (MSP) is a complete disaster recovery solution. That is seldom the case! Protecting your business against data loss and lengthy outages requires a holistic approach. The strategy needs to include planning, backups, snapshots, recovery, and testing. It means knowing how all the pieces of the puzzle fit together. Since 2014, companies have been increasingly adopting cloud backups. A 2016 survey by Clutch Market Research, found a surge in cloud-based backups, with 63% of businesses implementing them in the past two years. The same study found that most tested their cloud backups on a periodic basis. Still, many don't. Also, of those who test, most are only validating data backups, which won't help in the short term if your machine crashes. Given the costly nature of downtime, it's important you can restore your entire system, not just part of it.
The Metrics That Matter
Building your DR strategy correctly depends on first getting the right metrics in place. Before doing anything, you should have these two crucial targets defined for each primary system or application: Recovery Time Objective (RTO): The maximum downtime you can tolerate as your systems get back up and running again. Recovery Point Objective (RPO): The data loss you can live without before it impacts your business. The RPO deals with the last point in time to which you can successfully recover. Depending on the situation, an RPO usually varies from "almost immediate" to hours. RTOs and RPOs drive how often you need to perform data and system backups, and how long it takes to fully restore. The smaller these timeframes are, the costlier and more complex the DR solution must be.
Cloud-Based Backups a Great Start
MSPs such as Cartika offer Backup-as-a-Service (BaaS) along with Infrastructure-as-a-Service (IaaS). Cloud-based data backups have a number of advantages over in-house alternatives. They often come with built-in encryption and self-healing features. They also have better scalability and automation. If a vendor runs in a self-service model, customers can change features like update schedules and storage pool sizes remotely with a few clicks. Likewise, most systems come with a management console of some sort to simplify the process. You should find out the level of granularity the backups can support. Cartika's are very granular, right down to the file level. Every MSP is different, however, and if your data can only be restored at folder levels or in even larger chunks, you risk losing the many changes made since the last cycle. Despite their many advantages, cloud-based backups on their own don't complete the picture as they only capture data and not the Operating System (OS), databases (DBs), applications or settings. You need another process that lets you rebuild machines at the systems level.
Snapshots Complete the Puzzle
Snapshots or images as some call them, are required for a complete DR solution. These are often confused with backups, and many vendors help this along by using the two terms interchangeably. Taking initial, and ongoing snapshots play a vital role in most DR strategies. As mentioned, data backups in most cases, are not enough. A snapshot's job is to capture a system's entire hard drive or SSD at a given point in time and store it somewhere else. These are exact copies of the system, including the OS, DBs, applications, and all data on the drive when a snapshot is taken. They are key to fast RTOs because they allow you to quickly reset your system to a particular state before a failure without having to reload applications and configure system settings. The downside is that snapshots are storage and resource-intensive making them impractical for most companies to use as the only backup mechanism. That said, even if they are used, but not often, you can significantly slow down your recovery process - especially if you have months or years of OS and application updates to finish first. Read this article reviewing snapshots for more detail.
The Bottom Line
Buyers need to know what type of backups a vendor provides, how they work, and how to set them up to meet their needs. Preferably, this should be done before buying the service, and it should be based on solid planning. It doesn't stop there. You should regularly test data backups and the restore process. It's easy to kill two birds with one stone by using a snapshot to scale out another VM and test your restore process on that machine. Exclude lists also need updating over time to avoid needlessly backing up data. Disaster recovery is a dynamic and ongoing exercise that takes continual effort and refinement. Cloud-based backups are critical components, and so are snapshots. By combining both you end up with a cost-effective way to meet the vast majority of RTO and RPO requirements.