Data backups are – or certainly should be – an important part of your company's ongoing IT activities.
Local backups continue to have utility, for companies who may need fast restores of large data sets, or of system images. Originally – and still, for a surprising number of companies – off-site backups were done using physical media such as tapes, removable hard drive cartridges, and external hard drives.
Common sense dictates that data backup media should kept off-site. If something happens to your primary storage, that event may also whack any local backups or your ability to get to them. (Cue the fire engines.) It may be necessary, too, to comply with government or industry regulations, such as those for health care and financial services, and it’s part of any business continuity and disaster recovery scenario.
But transporting backups off-site via physical media has many points of failure. The media can be (and too often are!) lost, stolen, misplaced, or damaged. Or the data may be scrambled, e.g., by proximity to magnetic fields. Even when everything goes right, the time to prepare, pick up, transport, receive, and file the backup creates delay, which impacts the timeliness of any restore requests.
Online backup, using the Internet or other network connections, has been around for a long time. Some solutions used the network as online transport, writing the received data to tapes or other removable media. Others, under an evolving set of names for the service or provider like Application Service Provider (ASP), Managed Services Provider (MSP), Storage Service Provider (SSP), etc., kept the backup online, in Virtual Tape Library or other formats.
Enter Cloud Backup
Along with cloud-just-about-everything-else, the past few years have seen the emergence of cloud backup: online backup companies using cloud storage providers, and cloud storage/compute providers adding backup to their service portfolios.
In part, cloud backup simply means more use of third-party data centers – as well as greater storage capacity, better prices, and pay-as-you-go scaling.
Cloud backup shares many similarities with other types of backup, and most of the questions you need to ask a software vendor or service provider are the same. For example: Are files backed up whenever a change is written to the disk? What kind of backup is done: incremental or full? Is versioning available? How much does it cost to put a large restore on a disk or NAS and courier it next-day? Who has access to encryption keys and account access codes? What service level agreements (SLAs) are promised or available?
But cloud-izing backup has added more function-and-feature opportunities – and a few more concerns as well.
But First, A Brief Reminder: Backups Aren't Mirrors
Backups may provide some functions of live "drives in the sky," but it's important that you – and your users – do not confuse backups with live copies of your data such as RAIDed disks in a network-addressed storage (NAS) or array, or mirrored systems.
Mirrored or replicated storage is live data that is maintained in more than one location. That’s a good thing in the case of hardware failure (e.g. if a hard disk dies, you lose essentially no data) but if a user makes an error (such as deleting the accounting files) or the system is compromised (such as a from a virus) those unwanted changes are propagated.
Backups provide you with a way to recover from these glitches: retrieve a mistakenly deleted file, or go back to a previous point-in-time version, and so on. How much data and how fast it is made re-accessible are known as Recovery Point Objective (RPO) and Recovery Time Objective (RTO).
De-Silo-Ing: Expanding Who, What Can Access Backups
Cloud backups, like other cloud-based services, can be more flexible than a classic online backup.
"Traditional online backup has one 'in' and one 'out' – the user does the backup, and may access it," says Michael Fisher, co-founder and general manager at ElephantDrive. "The cloud offers a plurality of touchpoints and use cases. We and several other cloud backup providers expose access and other manipulation to the data both directly to end users and through APIs. For example, you can use the service for elastic 'overflow' to retain data for 30 or 60 days if you've run out of local storage capacity."
Symantec's Backup Exec, an on-premises product, lets users connect to cloud storage, says Amit Walia, VP of Product Management for Symantec's Information Management Group. "We have Open Source Technology plug-ins that work with cloud providers, such as Nirvanix and let customers write data to storage in the cloud."
Perhaps most significantly, backups can be used by Disaster Recovery/Business Continuity-in-the-cloud offerings from the same or other providers, holding virtual-machine images of company desktop or server applications, letting IT avoid the delay, effort and cost to maintain or quickly re-provision suitable hardware along with the OS and applications.
A growing number of companies, like Axcient and Geminare are offering DR/BC virtual-machine provisioning (assuming that a VM image has been previously uploaded), in concert with or as part of a cloud backup service.
And backups, or backup services, can also be used for online storage activities like file-sharing, sync, and other collaborative activities.
Axcient CEO Justin Moore says, "Because the cloud can be a seamless storage layer, you'll see more blurring of data management, collaboration, storage, and backup."
"We typically see customers who engage with cloud services like ours start using them just as backup, but evolving into using it as a productivity solution," says ElephantDrive's Fisher. "For example, you can give remote sales offices access to this data, avoiding the need to work through the company firewall. And if you have a video or other large file you want to share, you can do managed file transfers."
As with other cloud services, you probably expect that cloud backup means no hardware or software to install, no upfront or hidden costs, and easy web-based management. In reality, "no software" may still involve a brief script or shim loaded on each protected machine by the cloud backupware.
Don't take this for granted, though.
"You need to confirm whether this is a turnkey service or one that you have to integrate," says Greg Schulz, founder and senior analyst of the StorageIO Group. "For example, when I use JungleDisk, they provide software, which I install. I don't have to also buy Tivoli or Symantec Backup Exec, etc."
Location, Location, Location
With an online provider using its own or a single third-party's data center(s), you knew where your data was going. But a cloud backup provider may have multiple data centers, or work with multiple cloud storage providers, who in turn may have many locations. Autonomy (owned by HP) currently has 26 data centers, for example.
In some industries, there are legal requirements that data can't be stored in other countries, or possibly even traverse national borders. Those rules may include backup/redundancy sites that the storage provider uses.
Look for data locality assurance, says Richard Bojanowski, owner of X-stream.ly, a hosted service for building real-time applications. "The customer should know physically where the backups are located (geographically) and be assured any replication remains within a known geography to ensure data compliance and discovery regulations aren't violated."
Value-Added Backup Analyses
The big challenges of backup – or primary storage, for that matter – include finding what you want quickly, and finding patterns or other valuable information that are in the data as a whole.
According to David Jones, CEO, Data Protection and Emerging Technologies, Autonomy, the company's Intelligent Data Operating Layer (IDOL) software lets customers understand the meaning of data, including audio and video. This, says Jones, "lets companies make better decisions about placement, frequency, and retention which, in turn improve RTO and RPO and reduce backup windows and costs."
Opex Vs Capex, Volume, Etc.
Here's a few more cloud backup issues and benefits that your business should consider:
OpEx Versus CapEx: All cloud backup vendors tout the fact that the service is pay-as-you-go (and "pay-as-you-grow"). That can be gentler on budgets than are capital expenditures for storage hardware and software, plus service contracts and other costs. However, be sure to read the contract carefully for one-time start-up fees, and costs for restore requests, reports, or other activities.
Are public clouds a match? For smaller businesses, backups relying on public cloud storage can be a good match, minimizing local IT requirements of gear and staff. But for others, depending on amount and type of data involved, using the public cloud, or backup providers that use the public cloud, may not be an option. "A lot of data is so high-volume that network transport isn't feasible," says Symantec's Walia. "And security, scalability, and control are always concerns. Plus, moving custom data to the cloud isn't always doable."
Cloud Backup: Part Of A Total Backup Strategy
Cloud backup has a lot to offer, but that doesn't mean your company necessarily should go all-cloud. Increasingly, local backup software and appliances are including cloud backup, either as bundled services, or the "hooks" to let them easily connect to leading cloud backup and storage services.
For example, "NETGEAR's ReadyNAS storage line includes several implementation of cloud computing/storage features, including cloud backup," says Drew Meyer, senior director of product marketing, Commercial Business Unit, Netgear. "NETGEAR's ReadyNAS Vault automatically moves data from the ReadyNAS to a third party data center in the background. And for companies with multiple ReadyNAS appliances in their network, we also offer ReadyNAS Replicate, which lets them use their own gear as a private cloud backup target, taking advantage of capacity they've bought but aren't using."
Autonomy's services include a premises backup solution that gets its data from the cloud backup, based on analyses using Autonomy's IDOL to determine what data is most likely to be needed from a local backup.
And Some Backup Advice Doesn't Change
While cloud backup can bring better pricing and more flexibility, some aspects of backup still haven't changed.
"Look for clear specificity, e.g., for SLAs and for restore times," says StorageIO's Greg Schulz. "'Best efforts' is vague. Make sure that restore times are 'transit time via network or physical media to the designated site,'" not just for creating the restore image or media."
"Make sure a cloud backup provider has compliance certifications, privacy policies, and strong security," adds Scott Lee, director of marketing at EMC's MozyPro online backup service.
"As with any service, with the cloud you should always make sure that you know what you are paying for and what measurements are in place to show you are actually receiving the service," says X-Stream.ly's Bojanowski. "Ensure that you understand the service levels you can expect for transaction response times, data protection, and speed of data recovery."
Also, determine how the cloud backup service works with your existing backup tools and procedures. "Bigger environments typically have an investment in backup software and procedures," says NETGEAR's Meyer. "In moving to cloud backup, IT here is more conservative. They want to complement what they've invested in and ease over to newer more efficient technologies." There's also a financial aspect to this, Meyer notes: What's the impact of your licensing schedule for the backup software you have in place or your contract with a tape management provider?
"Think about what model you want to use," advises Meyer. "Are you ready to outsource all your backup data management? Do you want to move to an outsourced model in slow, careful steps, via a hybrid local/cloud route? Do you prefer to stay hybrid for security and performance reasons?"
And here's one final thought, from X-Stream.ly's Bojanowski that’s always worth repeating: "For a service that will surely be critical to your company, the best advice is to ask a lot of questions and get all commitments in writing."