Merylee,
Permit me to shed some light on the inclement weather issue, if I may. The tolerance of an infrared system such as the one being discussed here is indirectly proportional to the distance of its end points, AND the bit rate being supported over it.
As an example, allow me to again describe a situation I was once exposed to, about fifteen thru twelve years ago ('85 thru '88). In this situation multiple buildings were being served via T1 infrared "window shots in the canyons of Wall Street, in downtown Manhattan.
[[For those who've already had to endure my exploits in this area can simply click next, or hang in there with me for a few moments... and hopefully offer some comments and corrections, later on.]]
These beams of invisible light supported encrypted T1 links at 1.5 Mb/s carrying treasury applications between four buildings, as follows.
All of the outlying buildings were aimed at the hub. The "hub" building, where the mainframe was located, had all of its "guns" (i.e., transceivers) situated on the 8th Floor.
The shortest link was to west of the hub building, across a narrow one lane street a couple hundred feet away. It's guns were also on the 8th Floor.
The intermediate distance link was aimed north about ~ 700 ft, with the apparatus in this second building on the fourteenth floor. And the third link which was the longest of the three and to the east of the hub, was about 1400 feet.
Since the shortest distance building building was immediately across the street (a one way narrow lane) it effectively relegated both of the end points on this link deep into a canyon-like environment. Visibility was always clear between these two points, in other words, and even during the worst weather conditions the intrinsic quality of the data link was for all intents and purposes, unaltered.
The 700 ft building was accessible through a wider clearing on the northern side of the hub building, making it more exposed to the effects of weather, but still close enough to escape nature's wrath during most downfalls and even moderate levels fog.
The longest path presented almost all of the weather-induced problems, due to distance, its altitude (more likely to be hit by morning fog), and the angle at which the guns were aimed (due to the greater disparity in their respective floors, compared to the other two which were on the 8th and 14th floors).
What happened when inclemency occurs? This probably varied with each installation at that time. As a rule, when increasing thresholds (as the weather condition deteriorates) were met, link errors began to manifest. In order to mitigate the effects of errors (which usually result in time-consuming retransmissions), it became necessary to decrease, or downshift, the link speeds until an acceptable error rate was achieved.
More specifically, when problems stemming from weather arose the network control center had standing orders to reduce the link speeds in descending intervals of n * 64kb/s, until the links stabilized at some level where error rates are again acceptable. All the way down to 56 kb/s or lower, if necessary, and if the added time necessary for job completion permitted.
If the inclement weather remained at a constant but manageable inclement level, the reduced speed would usually work until the situation cleared itself up. If the weather condition continued to deteriorate, however, it became necessary to meet each level of severity with an additional downshift in link speed. In contrast to today's automated network health monitoring and provisioning capabilities, this was a truly manually intensive tracking and triage process.
At some point you reach a point of diminishing returns when the condition deteriorates dramatically, until at some point no decrease in speed helps the situation.
At the extreme, "zero visibility" conditions.. as you could well imagine.. are futile, altogether, right from the outset, and could only be dealt with by waiting them out until the fog or downfall lifted or subsided, respectively, or by reverting to some form of backup. If the application window did not permit waiting it out, during such rare conditions metallic T1s were used for restoration.
Ironically, the metallic T1s of that era were less dependable than these IR shots were, by and large.. this was prior to the time when fiber was used extensively for private line T1s, or just about the time when they were becoming made available by the competitive access providers (CAPs, such as Teleport, which was the only fiber-based CAP of that era), but that's another story.
The third building which was 1400 feet away had its guns situated on the 38th floor (a very sharp angle from an 8th Floor origination point under normal conditions, which was made less severe, however, by the distance of 1400 ft involved), and was the most vulnerable of the three to the effects of weather, fog in particular.
On very few occasions, relatively speaking, over a three year period that I can recall did we have problems with weather that could be classified as fatal. I can count on one hand where the entire operation had to revert to terrestrial backups. But there were many times when speeds needed to be reduced, rendering an impaired condition. In fact, the shortest path (directly across the street) would never present a problem. The 1400 ft path was almost always the one that would "act up" due to heavy fog and downfalls.
And whenever a level of impairment was reached, all links were affected equally due to the common equipment that was used to set the data rate at the hub site. Hence, the overall operation was vulnerable to the proverbial weakest link principle.
Today this would not occur, I wouldn't think, due to improvements in media access layer hardware, and autosensing techniques used in port adapters on switches and routers.
On the other hand, today's starting data rates would most likely be orders of magnitude higher than the T1 rates which I have cited here. Witness, the published multi-gigabit rates of Lucent's OpticAir and Terabeam's own device, which probably means that less severe levels of inclemency would cause problems for the same distances, and would spark the need for downshifting, sooner, rather than later.
Or, it could also mean that distances would need to be shortened across the board, or to a great extent, or some combination of the two. Of course, improvements in laser technologies and other areas of optics over the past ten years will more than likely offer offsets here, too, but just how much remains to be seen.
In fairness, I can only recall one or two occasions when all of the links had to be taken down completely due to the effects of weather. Again, however, I do recall running impaired quite a few times.
Also it would be wise to keep in mind that this was at a different time in history when data rates were far lower than they are today, and in the case I just cited only a single application was being supported. Work-arounds were easy to contend with, and in the worst case the metallic T1s, and even lower rated links, sufficed as backups.
The point to keep in mind here is that both distances and bit rate will play into the calculus when dealing with weather-induced impairments. And, of course, the need for unobstructed line of sight is an absolute given from the outset. If you can deal with the unpredictability of weather, and the 'given' that you have a backup for such contingencies, you may then find that window washers, hovering pigeons, ticker tape parades and trees growing in the path are less forgiving, still.
That last statement was an obligatory part of the prose which I feel compelled to present here, but granted, it's a superficial piece of trivia, for the most part. Unless...
FAC |
