Continuous Data Protection

The more technologically advanced we become; the more difficult it becomes to manage the data that technology produces. As computers increase in speed and storage capacity, the amount of data increases along with it and companies are struggling to deal with all of that data. Simply saving a daily picture of the data is no longer sufficient. The limitations it places on restorative functionality are tremendous.

Continuous data protection is a new concept in data backup. Traditionally, computer data backup is done periodically, usually once per day, though it can be done as infrequently as once per month. These backups generally saved files as they existed at the time of the backup and any variations of the files made during the day, week or month, depending on the frequency of backup, were lost with this method.

Computer data protection is going through a bit of a revolution right now. In the past, disk protection involved a periodic backup, usually onto a disk or tape recording system. This method has served its purpose well, but for all its benefits, it was still missing something vital. Continuous data protection involves continually capturing all data, including the changes made throughout the day, week or month, and tagging the varying versions for specific point in time recall if needed. Though disk based data protection can offer faster backup and restore time, it does nothing to help recover the data from earlier versions of the files being saved. Continuous data protection offers nearly infinite recover points so that any change to a document or file can be recovered and there is no requirement for the user or administrator to predetermine the amount or frequency of the recovery period. Continuous data protection is just that: continuous. The save points permitted by the implementation of continuous data protection occur each time a change is made to the data.

There are three different types of continuous data protection: block based continuous data protection, application based continuous data protection and file based continuous data protection. Each method of continuous data protection has its benefits, depending upon the situation and the user's needs.

Continuous data protection differs from traditional data backup in that the user does not have to indicate or mark a source in time in which to restore to in order to retrieve data. In other words, in order to restore to a certain point in time, the user does not have to save or mark a point to return to it. This is done automatically anytime a change is made. There are some programs that have been using this method for a while within their program, but they have been limited. Photoshop, for example, has a History function that allows a user to go back through the recent history of a document that is being worked on. The catch is that, how far back the program saves actions is limited and has to be determined beforehand.

Though some solutions are represented as continuous data protection, in actuality they only allow restoration to a fixed point interval, one hour, for example. While this is certainly more often than once a day or once a week, it is done automatically, rather than manually by an administrator. Continuous data protection is different from traditional data backup such as RAID, replication or mirroring in that these approaches merely protect against a hardware failure in storage by protecting the most recent copy of the data. Continuous data protection protects against data corruption and viruses and allows you to previous versions of the data.

Older methods of data protection employed a snapshot of the data. The snapshot continuous data protection takes of the file can fall into two separate categories. The snapshot can be a full copy snapshot that replicates the data set in its entirety. Often implemented in a process called mirror splitting, a full copy snapshot of a single terabyte of data consumes an additional terabyte of disk space.

A differential snapshot manages deltas that have only occurred since the snapshot was last taken. This method uses less disk space but requires more processing overhead. There are a few different methods for employing a differential snapshot, including Copy on Write, Redirect on Write and Write Anywhere.

The difference between these types of snapshot methods is in the cost of storage and performance. In addition, since the full copy snapshot does not rely on the main data store, they are usually more resistant to certain forms of corruption.

Whether the risk to data is from malicious code, such as viruses and Trojans, natural disasters that cause damage to hardware or simple hardware failure, today's companies face much greater risks to their data than ever before. Not only is there more date to manage, but also much less time to perform lengthy routine system maintenance. Today’s data centers are also less tolerant than ever to system outages.

In only the first half of 2005, there were more than 10,800 viruses and worms. This was an increase of 48% over the 7,360 documented in the second half of 2004. The increase in threat as well as the increase in variants is significant. This increase represents a new and distinct threat to systems which must be protected. The trend has also shifted from mass mailings to malicious code that is specific, modular and easily customizable.

Continuous data protection becomes vital to today’s webmaster. With continuous data protection, a website remains protected because the site’s data is continuously saved. Should something happen to the site, the webmaster can return to a snapshot of the site and restore the site from that particular point. Since the pictures are fluid, any previous modifications will be saved when continuous data protection is implemented. A webmaster therefore does not need to worry about losing the most recent updates to their site in the event that something should happen to the website or any of its pages. What’s equally nice about continuous data protection is that if a webmaster should make changes to a site and they find themselves unhappy with the changes, it is easier to restore the page to its former status.

At first glance, the two approaches to data protection seem similar. On closer examination, however, major and critical differences emerge. While the older version takes a snapshot of the data at a given point in time, limiting the user to specific way points in which to recover back to, continuous data protection provides a fluid and continuous picture of the files being modified, with no limitation to the amount of save points or the time frames in which the data is saved for restoration. There are three major methods of continuous data protection in use. Each offers its own benefits over the other two.

Block Based

Block based continuous data protection is a logical time-ordered cache of all block-writes across the storage network. Block based continuous data protection works without regard to the type of server or type of storage. Basically, block based approaches see every block-write move across the storage network and keep a time-ordered cache of those writes. Some block based solutions can instantly present a view of the disk at any time represented in the cache. Block solutions are perfect for presenting a view of a past point in time, but often require the user work to make use of that view. The user cannot access or restore specific areas within a document, only the full document or the entire disk. The benefits of block based continuous data protection include high transparency and no performance effect on the application being used.

File Based

File based continuous data protection is a capture file-system data and metadata events. File based continuous data protection run on the application's file server or workstation but present greater value since many applications and users make use of file-based data naturally. File based continuous data protection can have different policies per file or group, unlike block based continuous data protection can only be set per disk. For example, some files on a server may not need continuous data protection or another set of files might need a longer history of time captured. Restoring is also smoother with a file based continuous data protection, since you do not need to mount an entire volume view of some past point in time. The user can see individual instances of each file and choose the one they need. The benefits of file based continuous data protection include light weight, file-based policies and more natural recovery scenarios.

Application based continuous data protection operates from within the specific application being protected. Application based continuous data protection is similar to file base continuous data protection in that file serving is essentially an application. With application based continuous data protection, specific applications are responsible for taking care of the journaling necessary to recover to a point in time. Since it is tied to the application, application based continuous data protection can provide a far more detailed and richer set of recovery points. A database could recover a row or a column in a table from 2 hours earlier without interrupting the running application. With block based continuous data protection, the solution would need to present a view of the entire disk and application and would need to mount that view for use.

The objective of continuous data protection and recovery protocols is to allow storage pools to exist in one or more locations, presumably where a duplicate host can immediately resume operations with local data. Continuous data protection along with seamless data recovery is a very powerful combination that eliminates the lengthy and often-times frustrating restorative and replication procedures of the past systems. This combination can allow a company to recover business applications in a matter of minutes rather than hours and can restore to any point in time.

Choosing the right continuous data protection approach depends upon the user's specific needs for their data protection. If your data is file based, meaning it is strictly files and they are being used by typical office employees in creating and editing documents, then the file based approach would most likely be best for you. If your company mainly produces automated business applications such as XML packages, then file based continuous data protection would also probably serve best. This is especially true if you need to protect user end point systems like workstations and laptops. If your company is generally using a variety of applications, a block based continuous data protection might be the best option since the recall is not specific to a program, but a period in time. Finally, if the application your business is running has its own built in application based continuous data protection, then using that would probably be the best course of action.

One of the major concerns for the business professional is the protection of data while traveling. The time wasted when a document or file is lost and the time spent attempting to recover that information is a key concern for the modern mobile professional. In this age of laptops, the ability to backup and restore data is key to productivity. Relying on a system administrator to backup data might work in a traditional office environment, out on the road, where many professionals do what they do; this scenario is just not feasible. Traveling between the office and home, between airports or even telecommuting makes relying on others for backup impossible.

Business professionals today, more and more, need to be their own IT department. They need solutions for these issues that will save them time and increase productivity rather than costing them both of these. Some people might not be comfortable setting something like this up on their own, but in general, once it is in place, it is virtually automatic and end-user friendly. Continuous data protection is transparent and self-contained. It does all the work, leaving the user free to continue with their document of file. Any restoration point is automatically created within the system and requires no input whatsoever for the user.

Some data protection proponents recommend attempting to approximate the capabilities of continuous data protection by creating many snapshots. This is not a particularly attractive option as each snapshot is another object that needs to be created, bogging down the application itself and which is stored and managed at a cost to the company. Snapshot requires the administrator to predetermine restoration points before any failure event and then manage each point. Most continuous data protection solutions provide the ability to annotate a timeline with specific events or special times in the history of the data. This gives the user the ability to restore to any specific point in time.

Continuous data protection using real time data protection ad restoration saves IT and end user labor. It can restore to point-in-time and in multiple versions of the files retained. It can also provide data integrity in the case of a virus or some other type of corruption attacking the system. It allows the operators to restore files and applications to a point in time before the corruption once the problem is dealt with. Continuous Data protection backs up to a file server on a periodic basis and eliminates backup windows. It generally only backs up changed files rather than arbitrarily backing every file up whether changes have been made or not, making for faster save and recovery times. Most continuous data protection approaches optimize bandwidth and network transfer of data.

This data safety net provides real time backup for important information as well as continuous protection from viruses, file corruption or even the simple accidental deletion of a file, something everyone has done at least once in their lives. Simply restore the document from its last changed state, not its last saved state and continue as if nothing happened.

The key word in continuous data protection is continuous. Continuous data protection provides a continuum of restore points for a given set of data. This offers the ability to restore to any point in time, not just the specific times predetermined by a snapshot process. Continuous data protection provides infinitely granular recovery points and allows the restoration of applications to just prior to specific events. The restore point is selected and recreated after the catastrophic event occurs.

Continuous data protection does just that: it protects your business' important files continuously and in real time. These systems back up files at the moment they are saved, rather than waiting for a previously scheduled interval. It does this while still saving and cataloging the previous versions for instant recall later, if needed. The disk to disk storage capabilities are much faster than disk to tape approaches of the past. It helps control costs and increases user productivity. Valuable time is not wasted trying to restore lost files. It is easy to implement and it runs transparently in the background. No one will know it's there until they need to use it. Continuous data protection can protect key corporate information, even if it is stored on file servers and laptops.

In the past, companies were forced to choose between fast backup and quick recovery. This is no longer the case. With continuous data protection, administrators can create backups and restore to specific system recovery points quickly. With this system in place, in the event of a crash, critical servers are promptly restored online. This speed of recovery is critically valuable in today’s business environment in which companies must maintain data availability and minimize server downtime.

Continuous data protection is the newest, safest and fastest method for data protection. It provides transparent usability combined with unlimited recovery and restore points at any point in time, allowing instant recall of data should a system crash, hardware failure or any damage to hardware occur.