Usually about half way through each rotation I will try to find the Brigade XO and spend a few minutes talking to him about communications. By this time, the network is well-established, the unit has conducted a couple of CUBs, and by and large everyone should have a pretty good idea of how well we can (or can’t) communicate across the unit. One of the comments I frequently hear is that they “need more bandwidth”. I ask him why he says that and the normal answer is something about “well it just takes so long to download something and there is a big delay when we make a phone call”. This is usually about the time that I show them one of our bandwidth reports showing that in many cases, the unit has plenty of bandwidth to spare. This is where I being my discussion with him on why HCLOS is so important.
What is Bandwidth
Most non-signal people (and even a fairly high number of signal people) mistake delay for bandwidth and think that solving one will solve the other. For the sake of making sure that everyone is on the same page, a quick analogy:
Bandwidth is the size of the pipe. It is how much data we can push through the pipe at a given moment in time. Using the classic water example, we compare bandwidth by comparing a garden hose to a fire hose. The garden hose is able to move a fair bit of water out of the spigot, but the fire hose has a much bigger diameter and in turn, can move a lot more water through it at once. While bandwidth (or lack thereof) can play a role in the delay of data (if the pipe is full, the router will buffer the data or eventually drop it), this is not normally a huge concern.
Delay is the measure of time that it takes for a bit of data to move from transmitter to receiver. Going back to our water example with the garden and fire hoses, if both hoses are the same length and we turn them both on at the same time the first drop of water (data) will get to the end of the hose at the same time. On the other hand if we extend the length of one of the hoses and do the same thing again, the water will reach the end of the long hose after it reaches the end of the short hose. When we ping an IP address we are testing delay, not bandwidth.
We Need More Bandwidth
So this brings us back to the discussion with the XO and what a HCLOS link will do for him. Does a HCLOS link give the unit more bandwidth? Absolutely. Even a 512 Kbps link is technically more bandwidth although I don’t think that’s exactly what anyone has in mind. During the days of MSE, a 2 Mbps link was considered high speed. Now 8 Mbps and even 16 Mbps are fairly common (I’ve even read of 32 Mbps links being possible with upgraded HCLOS systems). This is some fairly serious bandwidth that in many cases can exceed all of the satellite bandwidth that the unit has. Unfortunately, when you come to the National Training Center because of a variety reasons, units are restricted to 2 Mbps on their HCLOS links.
So if that is the case is a HCLOS link still worthwhile? Most definitely. Even at only 2 Mbps a HCLOS link can be very beneficial to the unit. First, based on previous rotations few battalions will fill a 2 Mbps HCLOS link between them and the Brigade Main. We do see the link get saturated at times, but it is not the norm, and normally only occurs during heavy use periods (CUBs and things like that), and a good QOS policy (you have that running right?) will help take care of any problems this may cause. This means that during the majority of the time, there is bandwidth to spare going across the link. Instead, the unit sees the most benefit from the reduced delay that the HCLOS link provides.
When a packet of data is transmitted between two terminals over a satellite link, their physical location on the ground makes very little difference. Regardless if they are 5 feet away or 500 miles, the packet must first travel approximately 22,236 miles up to the satellite which it then bounces off of and then travels another 22,236 miles back to Earth to the distant terminal. Since the signal is travel near the speed of light (actually about 25% of the speed of light) it takes that signal between 300 and 350 milliseconds to travel between the two dishes. The return packet takes that same path in reverse that puts the full round trip at between 600 and 700 milliseconds. Compare this with a HCLOS link that is only travelling the actual distance between the two nodes (100m – 30 Km)who’s signal is still traveling at still about 25% of the speed of light and your round trip time is somewhere around 5-10 milliseconds.
Making Your Boss Happy
So this takes us back to making the boss happy. When we talk to the XO about how communications is working, really he is probably concerned about delay and not bandwidth. You can increase bandwidth all day long, and there is a decent chance he will not notice the change at all. Reduce the delay on the other hand and they will see an immediate improvement. The other thing that this brings up is how important it is for us to educate our boss on what exactly is going on with the network. We need to find out if we are meeting their expectations, and if they have concerns, figure out what those concerns are and how we can alleviate them. Do we have a bandwidth problem? Maybe, but probably not and the only way we can know is if we’re actually looking at how much bandwidth we’re using across the network, but that is another post.
One Response to “Scotty, I Need More Bandwidth!”
I love the idea of HCLOS in concept (low latency, high bandwidth). And it can even act as a backup in case a STT decides commit suicide (the router should just hiccup for a second and then start sending packets down the slightly slower pipe out to the other ends connection to the GIG). The hardest part is getting the BN’s to buy in and start using it (since there may be some tweaking of TOC location and adds a 15m mast to the footprint (I have noticed it is often the XO and/or the SGM who object to the introduction of an antenna like that, stating that is not neat addition to their orderly TOC area and that it isn’t tactical (some mumbling about it not being camouflaged (because a BN TOC isn’t readily apparent to any moderately sophisticated aggressor via optical, IR, EM or even noise (smh)))). That and getting the right people to the right location to get it set up and operational (since all the operators (and equipment) at the BDE level are tucked in the Signal Company).