Open-source products have always had a place in enterprise IT. Administrators and engineers usually have some open source tools around, whether on production systems, in a test and dev lab, or in an skunk works playground where folks can knock around ideas and try new things. Consider how open-source alternatives have reshaped the Network Management System (NMS) space.
Not so long ago, if you wanted to get a taste of a truly robust enterprise monitoring solution, you had to drop serious cash on the table--even if you only wanted a few of the features that the big names included in their packages. These days, IT pros have robust options from Nagios to Zenoss and others that provide key capabilities without the costs of commercial software.
So what about WAN acceleration and optimization? For many, WAN acceleration is still an untried technology. The skeptics that have no WAN acceleration now are interested, but need to be won over. With budgets tight, many IT managers want to drive before they buy. Unfortunately, shaking a demo appliance out of a Riverbed or a Bluecoat isn't always feasible, especially if you're a small to medium sized shop.
Enter open-source WAN acceleration. Don't expect to see open-source alternatives as mature as those in NMS, but projects like WANProxy and Squid can perform surprisingly well if you have some Linux talent on staff. WANProxy accelerates and compresses TCP and Squid does the brute-force work of caching; drop the software onto some spare servers and your staff can get a feel for the benefits of WAN optimization for short dollars.
I'm making the disclaimer up front that WANProxy and Squid are not generally scalable replacements for large-scale acceleration, but if you have the time, and you're curious about what sort of performance benefits might be out there, give each a try. If you've seen or had success with any other low cost or open source WAN acceleration or optimization alternatives out there, please share them with us! We're always looking to test out new and innovative solutions in the Network Computing labs.