Online Backup Best Practices – First in a Series of Five

Rob Cosgrove, CEO Remote Backup Systems

This five part series by Rob Cosgrove helps Managed Service Providers and Online Backup consultants design correct backup protocol for their customers.

This first article concentrates on optimizing backup time by identifying data that needs to be backed up online and excluding data that does not. This series also includes details on designing backup strategies, selecting backup types, defining schedules, and separating files by type into different backup sets.

Identifying Data for Backup

Data can be categorized into three tiers: critical (integral to ongoing operations), important (valuable but not mission critical) and archival (required but not typically retrieved).

Usually, the customer just wants everything backed up. Of course this isn’t possible at a reasonable price with today’s technology. Most hard drives contain many gigabytes of data, most of which is non-critical and can be recovered from original distribution CDs and other sources. Computers are not yet fast enough to prepare that much data to transmit over the Internet in a reasonable amount of time, and upload bandwidth is inadequate.

Here’s an example. The computer I’m using right now contains 43.3 gigabytes of data. My Internet connection is fairly fast – 12 Mbps downstream and 3 Mbps upstream. So, assuming 100% efficiency, I can theoretically send all this data up to the Internet in 1 day, 8 hours, 4 minutes, and 26.67 seconds.

That doesn’t sound too bad. But, no Online Backup Service Provider is going to take my connection at 3 Mbps, and I’ll never achieve 100% efficiency. The best I can hope for is an overall bit rate of (maybe) 900 Kbps, which is going to take 4 days, 10 hours, and 54 minutes. That’s too long.

Most businesses have an upstream bandwidth less than 900 Kbps. 512 Kbps is more common. That would take twice as much time.

In addition to the time it takes to do a backup, we must consider the time it takes to restore, called the Recovery Time Objective (RTO). This is a more important number because it directly impacts a company’s down time in the event of a disaster.

What is the acceptable amount of time between the loss of data and its recovery? If there is too much data (because of bandwidth) to download in an acceptable period of time, the Service Provider may have to deliver the restored data on a USB drive. This is where physical proximity to your clients is important. You could overnight a drive if your customer’s RTO is twelve hours, and depending on what time of day the data loss occurred.

You can test your upstream and downstream bandwidth at the following URL:

www.dslreports.com/tools

Then you can use one of my handy dandy online calculators to figure out how long it will take to do a backup:

www.remote-backup.com/calculators

 Clearly you have to be selective about which files you back up.

 Include & Exclude

Online Backup software contains rules that govern which files are backed up and which files are not. In RBackup and Mercury files and folders can be included and excluded from backups many ways. I won’t go into their specifics, but I will explain the basic concepts.

Global Exclusions

The best Online Backup software has features that allow you to globally exclude files and folders. The Global Exclusions list overrides all other file selections. For example, if your Global Exclusion list includes files that end in “zip” or that contain the word “cache,” files that match these patterns will not appear for users to select, and will be automatically excluded from any rule that might otherwise select them.

RBackup and Mercury both have powerful Global Exclusion lists.

A Global Exclusion list makes selecting files for backup far easier. For example, you could include the complete User Profile folder structure, and rely on your Global Exclusion list to automatically exclude specific files that never need to be backed up. It is a very powerful feature.

Following are some sample Global Exclusions used for Windows workstations. These are only examples. A typical Global Exclusion list is much larger than this.

Sample File Exclusions

*.sys	*.tmp	~*	*.dll	*.ocx	*.obj	*.temp	Ntuse*
*.log	{*	*.exe	*.msi	*.dbd	*.dsg	*.chm	*.pdf
*.ico	*.inf	*.js	*.jar	*.ex_	*.cab	*.bin	*.inx
*.hdr	*.cat	*.manifest	*.iss	*.fon	*.ttf	*.bak	*.zip
*.rar	*.pspframe	*corel*	*.txt	*.swp	*.dct	*.iso	*.bkf

Sample Folder Exclusions

Windows	*recyc*	Temp	*temp	Msocache	Symantec*
Installshield*	Webroot*	Cookies*	Adobe*	Mobilesy*	*itunes*
*cache	*search*	*recent*	*logs*	$recycle	Programdata
Statedata	*Norton*	Windows.old	*userdata*	Networkservice*	Localservice
.temp*	MSN*	*yahoo*	Apple*	Java*	Uninstall*
Winrar	Xerox	Sigma*	Vmware*	Windowsupdate	Nethood
Citrix*	Logitech*	*quicktime*	Corel*	SQL_LOG_Backup	{*
Downloads	NVIDIA	$*	RBS*	ATI	Google

 Sample Virtual Folder Exclusions

Windows	Temporary	Recycler
Application Data	Cookies	System
Temporary Internet Files	History	Nethood

It is also a good idea to exclude files based on attributes other than their names or folder locations. Depending on your service level agreement, you may want to exclude network drives, USB drives, CD Drives, Floppy Diskette Drives, and Memory Sticks.

It is also common to exclude files by size. Many Online Backup services exclude files bigger than four gigabytes, except for certain databases like SQL Server, Exchange, and Active Directory.

Read the next article in this series here.

Rob Cosgrove is the President of Remote Backup Systems, founder of the Online Backup Industry, and a vocal advocate for maintaining the highest standards in Online Backup software. His latest book, the Online Backup Guide for Service Providers: How to Start and Operate an Online Backup Service, is available online now, on Amazon.com, and at bookstores.

Remote Backup Systems provides brandable, scalable software and solutions to MSPs and VARs enabling them to offer Online Backup Services.