news / tech talk

DTNs

by Lee LeClair
11/10/2009
As seen in Inside Tucson Business

If you are a science geek, you may have read about a technology called Delay or Disruption Tolerant Networks (DTN). It is an interesting technology designed to meet a particular set of goals, but has caught the interest of some communities who see a different potential use for it. This may or may not work out for the folks who have high security hopes for it compared to the conventional Internet.

First, let us examine what DTN is and why it was developed. A prototypical use of this technology was established and implemented at NASA to meet some peculiar requirements they had. When communicating across very long distances to spacecraft and orbiting bodies, one can imagine that data transfer is not as casual as it is for the average Internet user. Due to these issues, engineers had to be careful about timing and transferring data during windows of opportunity when data could be transferred. That kind of planning and timing is manual and labor intensive so they conceived of a store-and-forward type of protocol that would ensure that all the data sent or received would automatically just "get there" whenever it could. To support that, the protocol relies on total buffering of all the data sent to nodes where it resides until it can proceed to the next node or final destination.

Again, one can see the value of that kind of architecture for the intermittent nature of data communication in space. A similar but much simpler store and forward technology that most people are familiar with is email, or more specifically the Simple Mail Transfer Protocol (SMTP). With email, you send a message and your mail server sends it to the destination immediately if it can. If it cannot send it right away but knows the destination exists, it stores your message, waits, and tries again periodically. It typically tries for about 3 days (depending on your server configuration) before giving up and notifying you that it could not send your message for X days. This is a very simple model that does not rely on multiple node hops like DTS but the principle is the same.

Other than the space communications use, one could see the utility of this kind of capability in other circumstances where communications cannot be relied upon to be up all the time. Examples might include military units in tough battle situations, climbers of Mt Everest, deep undersea expeditions, extended deep cave explorations perhaps. Anything in which communications might break down for a while, come back up for a time, then go down again. The technology is opportunistic about when connectivity is available.

The downside is that the nodes, whatever they end up being, need to be capable of storing fairly large amounts of data depending on how long communications are down. That implies that to build the infrastructure right, one would need to know or make assumptions about how long communications should or could be down and how much data nodes could store before they became overloaded. Another key point is that the current set of applications would need to be re-built to use the new protocol so one would need a rebuilt browser, email system, etc. to use this protocol.

Finally there is security. I mentioned at the beginning that there are some groups out there that believe DTN might be the solution to many of the security issues that plague the Internet today. From what I have read and seen, that was never the intent of DTN and it would fail at that goal. The intent was to create a very reliable method of store and forward data communications. I have not seen greater inherent security in this protocol than there was in the original Internet protocols (TCP/IP). Like TCP/IP, the intent of DTN is to provide a functional and reliable means of communication to further scientific endeavors. It will probably succeed at its goal and if it is co-opted for other purposes and comes into wide public use, it will probably be subverted by a subset of the population, just like the original Internet protocols. In evaluating any technology, always beware the mythical silver bullet fix to all your problems.

Lee Le Clair is the CTO at Ephibian. His Tech Talk column appears the third week of each month in Inside Tucson Business