IBM Corp., Microsoft Corp., Symantec Corp. and others have recently announced new cloud data storage services. These come on the heels of earlier announcements by companies including EMC Corp., Iron Mountain and Seagate Technology, to name a few. Despite pronouncements by some that cloud backup services aren't just marketing hype, serious doubts remain.
The good news is that to a limited degree, cloud backup is real and can save you money, when used appropriately. However, cloud backup isn't a magic bullet, and there are several issues to consider.
In this tutorial on cloud backup, you will learn the pros and cons cloud backup services, how cloud storage is changing data protection and disaster recovery and the differences in the cost of cloud backup vs. traditional backup.
Cloud storage means different things to different people. For some, it means easy access to secure, inexpensive storage from anywhere. In some cases, these are the same promises provided by past storage buzzwords, many of which failed to live up to their hype. During the last decade we have all witnessed the promise of the Storage Service Provider (SSP), then grid storage, then utility storage as being the next technology that solves all our problems. Here we are today with the next banner being waved, claiming that cloud storage is the answer.
There are several companies providing different services, all using the term "cloud storage," although they differ in many respects. The majority of these offerings are a variation of online backup services. Although there are other services being offered using the term "cloud storage," many of these are not yet practical.
Most business critical applications will continue to depend on access to high-speed storage. For these applications, any delay in access can significantly impact their performance. Additionally, network bandwidth over long distances will always cost significantly more than connections within the data center. Due to the laws of physics, long-distance communications take longer, and hence cost more than shorter connections. For these reasons it is impractical to consider cloud storage for many business critical data processing applications.
So, if cloud storage is practical, it will be used primarily for applications that don't depend on high-speed, high-bandwidth connections to storage.
This leaves open data sharing and data protection/data backup as the two most likely uses for cloud storage. Typically data sharing and protection are not as time critical as transaction processing. You are more likely to wait one minute to see a picture of your niece uploaded by your sister than you are to wait one minute to bid on a rare item on eBay.
According to major backup vendors, more than 75% of all data restore requests are for a single file, less than 30 days old. Thus, the issue of network latency will not be an issue for the vast majority of file restores. Restoring a few files over a wide-area network (WAN) will typically happen nearly as fast as restoring local files.
Also, a large and growing share of data resides outside of traditional data centers on a variety of devices including distributed on desktop and laptop computers. The exact percentages vary with the type of company, size and location, but one thing is clear -- the data being stored outside of data centers is an important issue.
Editor's Tip: Do you want to learn more about WAN optimization? Listen to our FAQ guide on WAN acceleration and optimization.
Conceptually, there is a large difference between protecting data, and providing disaster recovery. However, the way these capabilities are delivered is often very similar. Conventional data backup and recovery applications typically also offer a disaster recovery option, sometimes referred to as "bare-metal restore." However, disaster recovery involves much more than simply restoring an application server. Ensuring that the data is protected from natural and human disasters is an important part, which requires off-site storage of data.
For home users and SMBs, DR is less of a focus than is data protection. If a hard drive crashes, or a laptop is lost or stolen, the primary concern is being able to get your data back.
Also, many companies also choose to ignore the problem of distributed data by ignoring it, or by instructing users to "Back up their data to a network drive regularly." Those of us who have lost data probably do this, sometimes.
Both of these problems can be addressed with the current crop of cloud storage backup offerings available. By providing automated backup, data is actually backed up. By doing so to a remote site, a moderate degree of disaster recovery is also provided. The data is protected from local disasters, including fire, theft, floods, tornados and human errors by maintaining copies of files safely and securely off-site.
Editor's Tip: Learn about how other companies are using cloud services for disaster recovery. Read this article about how the USGA deployed IBM services for disaster recovery.
There are an increasing number of cloud storage (or online backup) services available. Currently there are approximately 40 online/cloud storage offerings available in the U.S. and Europe, with more coming online every month. Many of these are designed for home and small office users, although there are more than 15 offerings designed for corporate users as well.
The services designed for enterprises typically offer more features and control, along with a correspondingly higher price. If enterprise control is not required, and you have a lot of data to back up, one of the unlimited backup services such as Backblaze, Carbonite Inc., ElephantDrive, Mozy or others may be your lowest cost option.
Microsoft is even getting into the act. Microsoft has a Windows Home Server, which is really like a place to store information in your home. A Windows Home Server is also able to back up this data to another drive connected to the system. But if you want to backup your important data off-site for DR, you can set up your Windows Home Server to back up your data to the Amazon S3 service, which is a commercial cloud storage service offering from Amazon. So, for home users who want the ultimate protection from floods and fires, the cloud is the answer.
Editor's Tip: To stay up to date on the latest online and cloud backup news, bookmark our special section on remote backup and recovery.
When considering using a cloud backup service, it is important to look at a few factors. These are the amount of data you have to protect, the amount of available internet bandwidth and the amount of data that changes on a regular basis. The table below compares cloud backup vs. traditional backup.
|Backup factor||Cloud storage||Traditional backup|
|Amount of data||Best when the total amount to protect is less than 100 GB per 1 Mb of network bandwidth. For example, 100 GB can be supported by a 1 Mb WAN connection (such as DSL)||For large amounts of data, or for environments with limited network connectivity, traditional backup techniques are more appropriate.|
|Rate of change||Best when the rate of change is less than 10% of the total data per month.||For data that changes frequently, traditional backup methods that use local disk and tape, with tape transport off-site are more appropriate.|
There are a lot of cloud backup options available, and nearly as many pricing models as there are services. One of the cloud storage offerings that has received a lot of press is Amazon's Simple Storage Service (S3) service. This service is open to nearly anyone, with a set price for storing, uploading and downloading data.
Not surprisingly, many of the cloud/online backup providers use Amazon as the back-end for their service. Although the Amazon S3 service is not necessarily the least expensive way to back up a lot of desktop or laptop systems, particularly if you have a lot of data to protect. In that case, a service that does not charge by the gigabyte is often a better option. Some of the lowest cost options cost between $50 and $60 per year, which is equivalent to storing approximately 30 GB on Amazon's S3. So, if you are storing more than 30 GB, an unlimited account will save you money.
One potential issue is archive vs. backup. Archiving allows users to free up space by moving data off the primary system. However, several of the online backup applications are designed exclusively for backup. So once you remove the data from your primary system, it gets deleted from the online backup as well. If you want to store copies of your home movies for safe-keeping, it may be very expensive to store that 1 TB on Amazon S3. Expect to pay $100 to move it to the site, and then another $150 per month to store it. For archiving large amounts of data, the best option remains storing data on tape, or a removable disk drive for smaller environments.
Editor's Tip: Want to learn more about the differences between backups and archives? Read this article about archives vs. backups.
The chart below compares the prices of cloud storage vs. traditional backup.
Prices of cloud storage cost vs. traditional backup
|Backup factor||Cloud storage||Traditional backup|
|Initial Cost||Most prices range from $0 to $10 for initial setup.
Business accounts should expect to spend more.
|The costs for traditional backup can vary greatly.
Small business should expect to spend $1,000 or more
|Cost per gigabyte per month||Varies, prices range from $0 to $1 per GB per month for most services.||This depends on data retention amounts. Expect to spend about $1 per GB per month.|
Cloud backup is real, with services available from a number of companies. These services should be considered for any environment that does not already have a data protection plan in place. Many home users and SMBs rely on small systems for all their data, although they may not have an adequate (or any) means of protecting this data. Even many large companies who until now have chosen to ignore distributed data on laptops should consider using cloud backup services.
Russ Fellows is a Senior Analyst with the Evaluator Group. He is responsible for leading research and analysis of product and market trends for NAS, virtual tape libraries and storage security. He is also the primary analyst for coverage of selected open-system arrays and virtualization products. Russ is a well regarded and successful industry professional with 20 years of high technology experience, including product design, product development, systems engineering, business strategy development, competitive analysis and portfolio management within both the vendor and end-user groups.