Disclaimer: The United States Department of Transportation does not endorse any vendors mentioned in the following transcript
Q. Paul Olsen Asked: Are these losses cumulative total for the total distance?
A. Paul Olsen:
I think that was in reference to a particular slide. Maybe we want to try to go back to that one.
Pat Clair:
Sure. Just briefly, the further you go, the higher the loss numbers get. Your losses for antenna mount, your budgetary set aside for rain, those things are one time. When we go further, our losses are going to be higher. In that sense, the losses for distance are cumulative. The loss for budgetary allocations are not. If we have more than one obstruction that we have to traverse, those losses can be cumulative as well.
Q. Paul Olsen Asked: Fresnel zone is only relative to a parabolic antenna or yagi antennas as well.
A. Paul Olsen:
I don't know whether we'd agree or not agree.
Pat Clair:
Actually, I would disagree. With regard to Fresnel zones, the pattern that the antenna provides is going to spread. We don't have a perfect focusing that's available in the antenna technology that we typically use for these systems that keeps the signal completely coherent from point A to point B. We do have a Fresnel zone to deal with.
Q. Paul Olsen: From a performance and cost/benefit perspective, how would one evaluate an acceptable increase in noise in determining procurement options; are there guides or SW tools to determine this?
A. Pat Clair:
. Let me repeat the question, because that was kind of complex, then I'll proceed with the answer. If I understood the question right, how do we generate the cost benefit analysis relative to the noise in the system and the decisions that we make based upon that? While the systems, I think, are best engineered with a requirements-based approach that starts with what it is that you want to move from point A to point B. If you understand - if you needed to do just a small amount of data, a few bytes of data or if you needed 4 Mb, because you were doing full motion video, at let's say the MPEG 2 rate, the equipment that you're going to be using is very different. If you start with what it is you need to transport and you understand the nature of the link, you can very quickly narrow down the equipment that is available to you to be able to bridge that gap. You will have a cost number in very short order. Given the cost number for the equipment, that's the science of it. Now we move into the art of it. You have decisions to make about where you're going to put the receiver and the transmitter and how high and what are the obstructions are that you need to deal with. The manufacturer can help you with those things and give you some rules of thumb for the equipment that you have tentatively selected for that link based upon what you need to do, how far it is, and what you've got to send over it. You have a dollar figure and you know how much a couple of bundled T1s are or whatever the other costs might be to send that over a media-based system. You get down to a comparison of the two very quickly. That is essentially the genesis of having those numbers to be able to do the cost benefit analysis for system selection relative to the other alternatives that you need to be involved with.
Q. Paul Olsen Asked: If using 5.8GHz (28 dB) point-to-point antennae for a wireless bridge over open water, is degradation of the wireless signal ever of concern? Is degradation of the Fresnel Zone even of significance over open water?
A. Pat Clair:
Good question. The question is: Is Fresnel important over open water at 5.8 and 26 dB of loss? Water is an interesting thing. Water can hurt us and help us and it's kind of like counter-intuitive to how it does it. Water, when it's still, it's glassy smooth, can actually be more of an impediment than if the water is choppy. The reason is that the reflections tend to be coherent and they make some of the signal arrive at the receiver coherently late. The receiver gets bombarded by the original transmission and then the same transmission coming in some small portion of time later. If the water was choppy, there will be a scattering of the wave and it would not arrive at the receiver with sufficient energy needed to be accounted for. The water can be in the Fresnel zone if the antennae are too low, in the same way that land can be in the Fresnel zone if the antennae are too low. It might not be an impact to consider with regard to Fresnel, but it might as well. I think typically when we're dealing with water and radios; we typically place the antenna pretty high. With regard to some systems that span water, systems operators typically report having problems with the system when the water is glassy smooth, just as a rule of thumb and some of the physics behind it.
Q. Paul Olsen Asked: Could you expand on the ability to reach omni-directional coverage as a function of the frequency used?
A. Pat Clair: So the question is about the ability to do omni-directional coverage from a certain transmitter and frequency. A couple of issues, I guess. The omni-directional antenna will either be built in to the transmitter or they will be separated. I think for most of what we do in ITS, the omni-directional antennae are typically separate. They have system gains that are really well understood and typically the gains are pretty low, less than 10 dBm. The coverage is based upon two factors. We can talk about the radius of the coverage. That's based upon the power that arrives at the antenna and is focused by that antenna and is broadcast. Secondly, what are the engineering dynamics of the antenna itself? I want you guys to think of a bagel. The coverage of an omni-directional antenna, this looks like a stick just sticking up in the air. The coverage pattern of the antenna is a toroid shape, a bagel. The engineering that goes into the bagel determines how thick the bagel is and it determines how perfectly round it is on its outer edge. The engineering and the manufacture of omni-directional antennae is pretty good and well understood, but we don't get perfect bagels, like we do at Dunkin' Donuts or wherever you go. There's a jagged edge to it and signal does fall off once you pass a certain distance from the transmitter. I would encourage you to look at the antenna plot relative to the certain antenna that you're thinking about using and understand how thick or thin the bagel is based upon that plot as well. We will have a broadcast footprint, plus and minus a few degrees from the antenna. Then we will have a radius from the center of the antenna. Of course, the antenna sits at the very center of the bagel, the toroid shape.
Q. Paul Olsen Asked: if low frequency signals are more desirable, why would you choose GHz wireless signals for short distance applications? Basically, there are other conditions to see before making decision on low vs. high frequency even for short distance. High frequency has inherent benefits in some cases. Correct?
A. Paul Olsen
Basically, there are other conditions to see before making a decision on low versus high, even for short distances. High frequency has inherent benefits in some cases.
Pat Clair: The question is: How would you differentiate between the choice between low and high frequencies and what are the drivers for that? From a physics perspective, I wanted to provide information that lower is better. But we live in a regulated environment. The power and the coding that's available and the channel construction or allocation within the frequency band is set by the FCC. Some of the higher frequencies have more advantageous banding than the lower frequencies. Some frequencies are allowed to transmit at a higher frequency. Some frequencies are allowed to use certain coding technologies that others are not. Very quickly, you can appreciate that this is a complex choice, decision matrix that would need to be understood. You can compare the different kinds of equipment and look at how they're doing their coding and look at the launch power. Typically, you could understand that the manufacturer is going to tell you what the maximum launch power is, because that number determines how far the spanning capability is between point A and point B. They will tell you the highest number that they can use. They didn't pull the number out of the air. They used that number because it's what the FCC told them to use by regulation. In general, the physics tells us one thing and on the other hand, the FCC tells us to do something else.
Q. Paul Olsen Asked: Is the Fresnel zone 2D or 3D? Is it related to polarization of the signal?
A. Pat Clair: It is 3D. I have trouble showing something 3D on my slides with any degree of success. You can think of that thing as a football, as a rugby ball between transmitter and receiver. It is this series of conic sections that are tapered on either end that goes from point A, the transmitter, to point B, the receiver, and being fattest in the middle. It is 3D, very much so.
Q. Paul Olsen Asked: Are there loss tables for things other than rain, like for a dust storm?
A. Pat Clair:
There are loss tables for dust storms. With dust you can expect the worst circumstances when the dust is electrostatically charged or the dust is wet, moist or something like that. We have terrible results at high frequencies trying to go through fog. The higher the frequency, relative to whatever small portion of the wave is. You could think if it was a rain droplet, what percentage of that wave is that rain droplet going to nullify? If the dust is charged or wet, you could think about it in terms of the same effect.
Q. Paul Olsen Asked: What field method/equipment is typically used to determine that a suitable Fresnel Zone exists between two points?
A. Pat Clair:
All the manufacturers that I'm aware of can provide you with equipment or will do the frequency analysis for you that will tell you if the distance between A and B is suitable for what you want to do. Wireless is both science and art. What I've presented for you this afternoon is the science of it. The art is that there will be numbers that you can derive by formula, but sometimes you've just got to go out there and see. A signal will be broadcast and received at the other side. Crudely, you could look at it and you will know that it's going to be a problem because things that are in the way cause problems. You can understand the diameter of the Fresnel zone with that simple equation. From a diagnostic trace perspective, the manufacturers will have the frequency analysis tools to tell you if the Fresnel zone was good enough in order for you to send the amount of data that you anticipate sending over the distance that you need to send it.
Q. Paul Olsen Asked: I assume that the Rules of Thumb for maintaining signal are good for maintenance group. For design people, there should be design software tools that would help you review these rules.
A. Pat Clair:
Right. The rules of thumb with regard to losses that you should incorporate into this thing that we call the loss budget ought to be written down and developed for your system. Different places have different geography. They have different attributes that we might call them like urban development, light urban, suburban, or rural. Those are very special words, because they mathematically describe the cluster of manmade objects. Urban is not the same as desert or rural in that regard. If the manufacturer is using tools for you and you gave him a zip code or you gave them a series of latitude and longitude, they had some tools to understand whether that was urban, light urban, suburban, or rural. They can plug that into their tools and they will be able to tell you with some degree of assurance that they think that would work. Then there would be a caveat about it would be really good if you could take a picture standing at where you think the transmitter would be to where you think the receiver's going to be so we could see what's going on there as well, because the data driven mapping tools that are available don't take into consideration stuff that was built just recently. That's a little bit of a wild card. That addresses the art of it. With regard to the need to understand what your systems are, I would just encourage you to write everything down to completely understand the design process, what went into the link, what kind of power margin you have with regard to that link from A to B, so that you could understand the reliability and the quality of the data that you should be able to send across.
Q. Paul Olsen Asked: What about 802.11p and a discussion of DSRC?
A. Pat Clair:
Basically, and at a high level, sure I can discuss 802.11p for you. The IEEE 802.11p has been set aside to do short range data transmission in a vehicular environment. That's pegged to vehicle infrastructure integration using the dedicated short-range radio. We're talking about short datasets not going very far, an ability to set up and tear down communication paths pretty quickly, all while maintaining anonymity of the receivers and the transmitters. So the 802.11p is an ongoing effort right now. I would just say that you kind of have to watch this space. There'll be interesting things happening there with regard to 802.11p, DSRC, and VII. But not all of the answers are completely known yet, but when everybody looks at this, we see the potential for wonderful things related to safety along the highways and major arterials.
Q. Paul Olsen Asked: Isn't the 802.11n standard still in draft? What risks are there in using the standard while still in draft status?
A. Pat Clair:
Good point. The question was about what's with the 802.11n? How come people are deploying when it's still draft? The caller is exactly right. It is draft. But people are so hungry for data. Some universities are pushing forward with 802.11n deployments and they're going forward. I think the reason that they're going forward is that the frequencies and the physical capabilities of the radios are really well understood. The thing that allows them to go forward is that we have software that typically makes theses systems go. The oscillators and the physical parts are not going to change. The gamble is that we could deploy 802.11n and if we needed to change something, we could repump the firmware for these radios and we would be just fine. I do not know when the standard will be ratified. Lack of ratification is not keeping people from deploying it. I think from memory, you may be able to Google this or not, but I think Duke University and a place called Morristown College and there are some other universities that are deploying it. If you wanted to Google that, you could find some references for people. These are big IT shops. They know what they're doing and their students are very hungry for data. They see an ability for 802.11n to provide what their customers desire.
Q. Paul Olsen Asked: What are the units for "r" in slide 31 about Fresnel radius?
A. Pat Clair:
That's feet.
Q. Paul Olsen Asked: Isn't that relative to the distance between the two points?
A. Pat Clair:
Yes. If the distance is further, that radius "r" is going to be higher. As the signal spreads out, it will make that football bigger in girth in all three dimensions.
Q. Paul Olsen Asked: Can you please touch upon WiFi and WiMax with regards to what you discussed, and what it means for the transportation and transit applications?
A. Pat Clair:
The question was, just to paraphrase, what kind of impact can we expect WiFi and WiMAX to have on transportation? There is tremendous discussion relative to WiFi and WiMAX relative to transportation needs. It's all related to build and buy decisions for planning organizations' transit agencies. Let me expand it and throw 802.11e in there as well. There are services that are required by transit. Typically, they want to know where the buses are. They want to know the rate that the bus is traveling at and given location and rate, they can make some kind of forecast about when it's going to be at the next stop. It's very powerful information. That information about where it's going to be at the next stop could be broadcast to people that are frequent riders of the bus and could show up on their PDAs or on their cell phone screens. "The bus will be here in 2 minutes" or something like that. The capabilities to provide that, the database work and the ability to derive that information are well in hand. The caller's question about what is the best channel to do - that is up in the air. We typically don't know what the optimum way to do that currently is. There's been a tremendous amount of discussion about privately owned systems to do that. In the United States there's starting to be discussion about utilization of 802.16e to do that, because it provides us an ability to say that in this digital transmission domain, some packets of information are more important than others. A quality of service differentiation is going to be possible that will drive implementation of new features and services for people. That's a buy decision to use somebody else's network, whether it was Sprint or something from AT&T or Verizon. In the build decision, the transit agency would make a decision about let's say a WiMAX type of system. They would have to put up the backhauls and they would have to put up the connections that would broadcast to their subscribers and to take the data from the buses or the trains or whatever it is that they're using. There are really interesting questions that are coming up relative to build and buy and the capability of the different communications protocols to do that. We can do currently tremendous things. Of course, on private systems where you put it up and build it, you can have more control over the quality of the service and entitlement of the users. You don't have to get in line with anybody else that's using a common system. You have complete control over it yourself. I hope I've been able to answer your question. It's kind of ongoing and unfolding as we speak. I think that the introduction of 802.16e in the next year in selected markets in the U.S. is going to have a tremendous impact on this build-buy decision and may kick off an explosion in services and capabilities. We'll have to see. I'm not too good on reading the future yet, but we'll have to see how that unfolds.
Q. Paul Olsen Asked: How about interference and collision with all signals in the air? How about interference and collision with all signals in the air?
A. Paul Olsen:
I'll tackle this one to give you a break, Pat. Most of these systems are based on the Ethernet standard. The Ethernet standard was essentially a wireless standard stuffed onto a coax cable. Prior to that was the ALOHA radio network in Hawaii. It was all wireless and they dealt with collisions. You either ignore them or you use the techniques in the original Ethernet standard to deal with them. Let's go on to the next one.
Q. Paul Olsen Asked: Can full motion video be successfully transmitted over the half-duplex wireless links that are common amount many wireless manufacturers in the ITS market?
A. Pat Clair:
Can full motion video be handled over a half duplex? The answer is sure. This is an interesting thing. Given the Wisp model, a wireless Internet service provider or WiMAX, that model, almost all of the traffic is downstream with very little upstream. If you're going to send something, the needs get hungrier, wider. But mostly it's your keystroke traffic. You're perhaps downloading all those YouTube videos. In ITS, if we're talking about video, we typically would have cameras at the end of that. The model is backwards. We can, in the absence of a lot of downstream traffic in that model, utilize the channel capabilities for upstream, MPEG2 or MPEG4 or motion JPEG. Depending on system selection, there would be enough room to handle hundreds of kilobytes or 1 or 2 Mb or maybe up to 4 Mb for MPEG2. It gets down to several things. Can the system do it? Is it rated? Have you chosen the right distance? Do you have enough overhead to account for the fact that it was half duplex? When the manufacturer told you it could do 30 MB per second, was that the sum of the upstream and downstream paths? Can you tune that somewhat? If you can't tune it, have you exceeded 15 Mb? Yes, you can do it. You just have to look at the numbers and understand what you're buying. You may only be buying half of the capability of the radio, because your traffic is predominantly in one direction.
Q. Paul Olsen Asked: Is the bit transmission rate a direct function of the frequency range?
A. Pat Clair:
Let's tackle the first one about the bit transmission rate being a function of the frequency. It is a function of the frequency, but in a more targeted way. So within every frequency allocation, there's going to be subchannels that are set aside. The number that is typically in the industry for high data rate traffic is that you need to have 10 MHz. Let's just say, for example, if the signal starts at 5 GHz, you can have as many channels as you want that the FCC is going to provide you. But each of those channels should be 10 MHz wide for you to do let's say video. It can be smaller if you play some tricks and what not. But 10 is a useful rule of thumb. You might have 10 or 20 channels starting at 5 GB and then running up say, inclusively. If it was 10 channels up to 100 MHz of space. So 100 MHz added to 5 GHz. Depending on how the system coded the data, it may dedicate channels. It may lump them and aggregate them. It's really a question for what's the capability of the system based upon the coding? It's important. That information is on the manufacturer's cut sheets. It will lead you to the right decision.
Q. Paul Olsen Asked: What are the factors to consider when deciding to opt for a licensed link over an unlicensed link?
A. Pat Clair:
You know, I'm a requirements guy. Start with what you need to do and what your budget is for it and how long the link is going to be up and what your concerns are for it. If you take a requirements driven view, it will lead you to ask the questions properly with regard to the equipment that you're seeking. How far do you have to go? That will lead you to power. Depending on the power, it may take you away from unlicensed systems immediately. How much data do you have to shoot over the link? There are some unlicensed systems that will do something approaching 300 Mb per second and even more, but they don't go very far. Take a requirements driven approach and you get a very short list of what you need. Then it's a process of elimination. Start with the total range of things that are available to you. Do your requirements. Put down what you need and then you can start eliminating things. What you're left with is your decision.
Q. Paul Olsen Asked: How do wireless networks perform when integrated with a fiber network?
A. Pat Clair:
Typically, not a problem. With regard to the integration of wireless and fiber networks, you typically would have an 802 type Ethernet kind of a switch that happened to have a fiber interface, perhaps gigabit fiber. If you had an integrated wireless system, it would just plug into that in one of the ports. Your concern is how much data do I need to send over it and what are the bottleneck issues? That's typically a function of systems administration and your expectations about how much data you can do. With regard to the protocol management and the ease of making it talk at this basic level, those things have been pretty well established and the engineering is very well understood now.
Paul Olsen:
I think to add to that, there's some issues of latency there as well. It depends on where the wireless is in the fiber network. If the wireless is connecting a couple of fiber networks in the middle, you might have some issues you want to deal with. Again, it gets back to discussing your requirements.
Pat Clair:
Yes. If it's fiber wireless and fiber, you're going very fast, slower, and very fast. You can have latency issues because of buffer handling on either side of that wireless link. Think about, you know, there's time associated with downshifting and being able to send that data.
Q. Paul Olsen Asked: Are their good resources for the theoretical or practical calculations of distance, etc (i.e. any good spreadsheets)
A. Paul Olsen:
Pat, I think the best way to answer that one is on materials that we post after the session here. We could probably have a list as long as your arm on this one.
Pat Clair:
I'll tell you what. There's several good textbooks that are out there, but the best textbook that I have seen with regard to wireless is not a textbook. It's a book by Cisco and it's called Deploying Wireless IP Networks. It's written in the style that makes it very useful to people. I think that's exactly the name of it, Deploying IP-Based Wireless Networks by Cisco Press. Just to follow up, I don't know who the caller was, but if someone wants to call me, I'll give them the ISBN number. I just have to get back to my office to get it off the shelf.
Q./Comment Paul Olsen: It's about bits per Hz. Higher bands have fatter channels but you pay for it by having to be more directional due to losses inherent in the physics. Good engineering is the art of the most appropriate compromise.
Q./Comment Pat Clair: There was a question that came in late about is 900 MHz capable of handling streaming video.
A. Pat Clair:
The answer is yes, but you have to be very careful about the coding for 900 MHz radios. Some of them are frequency hoppers. That won't work. But if it's not going to hop, it will do streaming video.
