Commentary

File Transfer Acceleration: What Seems Good For You May Really Be Bad

Imagine. You like cars. Fast cars. Racing cars. And, one day, you're given the keys to an Aston Martin DB9. But it gets better -- you can drive as fast as you want. No speeding tickets, no traffic --in fact, no one else is on the road. But wait, it gets even better: there's unlimited gas. It's a six-lane highway and you can weave around as much as you want because, again, there's no one else on it.

Now, let's put some other cars on the highway, though, of course, they have to observe the speed limit while you do not. And, better yet, they have to yield to you no matter what lane you decide to use. Fire trucks? Ambulances? Perishable cargo? Hey, too bad. They're all relegated to second-class status behind you. Let them figure out who gets priority, as long as it's behind you, because YOU are priority #1. Simple example, right? But while it may certainly be exhilarating to drive unhampered, it does have ramifications for everyone else.

Because of the incessant market need to provide content to the viewer regardless of where, when, and how that content is consumed, all options for moving content quickly are being examined. Does it make sense to have someone carry two 50 GB Firewire drives on an airplane to move content from Australia to London for audio mixing on a feature film? The answer, of course is: it may. Especially if there is low network connectivity between the two locations. In that case, the issue is simple -- it's not the cost of the plane ticket, it's the time it would otherwise take that carries the larger cost.

advertisement

advertisement

Who Has Control?

So what happens when you are not the only one who can drive as quickly as possible on the highway? What is the practical, real-world result of enabling anyone and everyone in an organization to accelerate the movement of digital files that need to move from location to location? Sharing all the available bandwidth by every user who believes that their content should take priority over another user's content is a sure way for no content to get to its destination by the required time.

Uncontrolled point-to-point file acceleration is actually potentially injurious to organizations in the media and entertainment industry. Does the IT staff really want to preside over X number of federated point-to-point transfers and then query -- after the transfers have been completed -- as to what the activity was and what bandwidth was consumed?

And what about security? Do you rely on each group of users in each federation (e.g. each division of your organization) to use or not use, for example, media encryption? What if one group just doesn't encrypt the payload? How do you, if you are in the content protection group, deal with that? And, why would you leave those security choices in the hands of the users? Because, let's face it: someone WILL forget. And if that content, for that transfer, is stolen, who shares that risk? The user who initiated the transfer or the department responsible for the corporate network resources used to transmit the content?

And that's only one of the major problems with federated point-to-point file acceleration systems. Because while it may work great as long as not too many people are on the highway, it doesn't address which digital media files should take priority over others as determined not by the users, but by when the content needs to arrive at each location. The only real way to implement a scalable digital media distribution solution is to manage the distribution node centrally and apply business rules, policies, and priorities to the digital media content traversing both your corporate network and the open Internet.

Accelerating file transfers is not simply a technical exercise of TCP/IP window scaling or implementing UDP-based WAN acceleration. There must be in place a method of arbitrating how much bandwidth is applied to specific classes of content in order to ensure that the digital media supply chain does not go interrupted.

Next story loading loading..