Networking

WiMAX's slow rollout may be technical

There have been all sorts of possible explanations as to why WiMAX is not readily available. Some recent news may very well point to the real reason for the slow rollout. Quite simply, everyone is having problems trying to get "big enough pipes" to the cell towers.

There have been all sorts of possible explanations as to why WiMAX is not readily available. Most refer to business plans and political issues. Some recent news may very well point to the real reason for the slow rollout. Quite simply, providers are having problems trying to get "big enough pipes" to the cell towers. The InfoWorld article, "Backhaul woes slow Sprint's WiMAX rollout," mentions:

"What's holding Sprint back is simply the logistics of building the network, and specifically the problem of provisioning "backhaul" connections to the Internet. Mobile operators typically lease T-1 lines from their cell sites for these backhaul links, but those lines only provide 1.5Mbps. WiMax is designed to deliver more than that to every subscriber."

Existing cell tower bandwidth

Some history might be in order here. A typical cell tower that was in a fairly active area would have 3 to 4 T1s aggregate bandwidth. If that particular tower was upgraded to 3G then usually one or two more T1's were added to the mix, bringing the total bandwidth to approximately 9Mbps.

Bandwidth and range advertised by WiMAX

WiMAX has a theoretical maximum bandwidth of 75Mbps using 64QAM 3/4 modulation. WiMAX has a theoretical maximum range of 31 miles with a direct line of sight. It's not too hard to see why WiMAX backhaul may become a real issue. Requiring almost ten times the bandwidth that exists at each cell tower is a very significant increase. I also wonder if the increased range was factored in. A larger coverage area does not increase the bandwidth required per user, but it does increase the aggregate bandwidth as more users could be in each cell's coverage area.

What's the solution?

As a network engineer, increasing bandwidth to a facility is my bread and butter. A typical solution is to jump up to the next T-carrier, which in this case would be a T3. The bandwidth supplied by a T3 is approximately 45Mbps. One challenge with T3 cabling is that it's not just your simple two twisted pairs of wire anymore. The copper solution for T3 is a pair of coaxial cables -- one to transmit and one to receive -- with BNC connectors. Another challenge is that T3 signals can only run short distances -- standard distance is 380m-over copper.

Service providers are well aware of this and now have two challenges to overcome, increased bandwidth needs and expensive circuit extension challenges. There are two viable approaches being considered, and ultimately, it appears the answer will be a combination of both.

Since there is still some question as to what the ultimate bandwidth requirements will be, the logical choice is to just run fiber circuits to the cell tower. That will allow almost unlimited bandwidth, if implemented properly, by having extra dark fiber. There is major sticker shock with this approach, though. I have heard figures in the range of US$50,000 to $100,000 to run fiber to the cell tower, and that dollar amount is based on having a fiber network reasonably close to the cell tower.

A more logical and cost-effective approach would be to use microwave backhaul instead of fiber or even a composite of both. For example, one topology would have each remote cell tower connecting to a centrally located cell tower via an OC-3 microwave link. The distribution cell tower would be chosen or located so as to facilitate quality microwave links with all of the remote cell towers as well as the most cost-effective fiber run to the system's fiber network.

One interesting tidbit about this topic is that the rest of the world is significantly ahead of the U.S. when it comes to using microwave links for network backhaul. According to that same InfoWorld article:

"So the carrier wants to use point-to-point microwave wireless connections. Though these are used widely in other parts of the world -about half of all backhaul in Europe is microwave, and more than that in Asia, IDC analyst Godfrey Chua estimates-Sprint is practically treading on virgin territory by seeking it in the U.S., according to West. There are few engineers well-versed in setting it up, and Sprint has to overcome zoning issues for many installations, he said. The procedures of setting it up are also quite different, with the need to find an unobstructed line of sight through the air and deal with zoning issues."

Final thoughts

One has to wonder why this is even an issue. Isn't it logical to think that this should have been looked at right away? Is it possible that a technical oversight may be turning into a major business headache for Sprint and others? The recent interest of Comcast and Time Warner Cable in Sprint and WiMax, as pointed out by TechRepublic Executive Editor Jason Hiner in his post, "Sprint CEO stays the course on WiMAX, but launch won't happen in April" may be coincidental, but both players have access to significant fiber networks and that might help Sprint's cause. Hmmmm.

I also did a little poking around, and I see that Clearwire -- on and off WiMAX partner of Sprint -- is advertising for microwave engineers. In fact, there seems to be a significant number of openings for microwave engineers. Hmmmm.

About

Information is my field...Writing is my passion...Coupling the two is my mission.

15 comments
bobhaines
bobhaines

The question of mesh vs trunk is one of redundancy. Do you want the subscribers to lose service if one trunk goes down. Having a mesh network costs more than a single trunk. It is my understanding that the question might reduce to an economic one. It costs a fixed amount to set up the microwave towers and a recurring cost to keep the bandwidth flowing. The service provider has to absorb the startup costs until they have enough customers to cover these expenses. How long does it take before the ROI turns positive or does it ever turn positive. You need deep pockets to fund a venture like this.

Sensor Guy
Sensor Guy

Your comments on the T1-T3 issue are right on. Despite all the changes in telecommunications, most engineering departments in carriers are still dominated by management that came up the copper world. It'll be a while before Sprint overcomes this cultural problem. Although I have a background of traditional data and voice, convergence has given me the opportunity to move more into the non-wifi wireless world of late. I have seen the definite advantages the Asian and European networks have over us and I agree wholeheartedly with you that the backbone is beginning to hold us back. Towers, regardless of the wireless technology they are broadcasting, Wi-Max, Wi-Fi, MPT-1327, P25 must have a route back into the backbone. Other than using radio itself (in which case they are just a repeater) the tecnologies are traditional copper (T1/E1 and T3/E3), fiber (MAN/LAN or ATM OCx), Satellite (e.g. VSAT), and lastly microwave or light transmission (eg, infrared). Of all of these, the easiest to set up with the minimum of infrastructure for the longest transmission disatance is microwave. Nowadays, I recommend microwave as the initial tower backbone technology, then go for the fiber MAN/ATM if you can. Once you have fiber on the ground and microwave in the air, you have the best of all worlds.

andrenym00378
andrenym00378

Isn't a T1 Line still the best solution? T3 is just way to expensive!

panTribe
panTribe

1. Your 1st hmmm is unwarranted. Your 2nd hmm is on the mark, though. 2. I once worked with Sprint and produced an analysis in early 2007 on backhaul being an issue for 3G/Wimax, including what Europe had done differently to avoid this problem. It never went beyond my director's desk, but I enjoyed reselling my findings to outside interests who 'got it' after leaving Sprint Nextel.

Michael Kassner
Michael Kassner

Is it possible that the technical issues were not resolved before advertising the rollout schedule? Sprint already has fiscal issues, where will it get the finances to support WiMAX's real bandwidth requirements?

Michael Kassner
Michael Kassner

I'd have thought that this would have been all determined a long time ago. All appearances and information is coming across that the they have really underestimated the backhaul bandwidth. I maybe missing something but those types of calculations are used by countless engineers, including me on a daily basis. As for the mesh versus trunk I again agree. The only way it might be cheaper is if the cell towers are in close enough proximity to not require directional antennas, then a complete mesh could be a possibility. Even if directional antennas were required the horizontal coverage area could be sufficient to allow a partial mesh with more than one tower which would increase reliability.

Michael Kassner
Michael Kassner

I was curious to learn what type of equipment you suggest for the microwave backhaul?

Sensor Guy
Sensor Guy

There are many factors by which you can judge a telecommunications link. One of them is market presence, which usually is translated to more availability and "trust" in the link. T-1 obviously fits that well, since it is probably the current predominant mid-range speed link type among commercial entities in continental US. However, if I was building a backbone today, I'd ask if the backhaul link will be used for TCP/IP protocols. If so, I'd then lean towards some sort of business grade DSL, which can reach speeds of up to 3Mb/s on the same copper, depending on the carrier facilities available. If you can live with IP, DSL is a hell of a lot cheaper than T-1 prices. I also would mention that many of my friends, if they control both ends of the copper can cheat by making the T-1 an E-1 (European version of a T-1) which runs at up to 2.048Mb/s, the old version of the 32 circuit ISDN 64kb/s bundle.

Michael Kassner
Michael Kassner

I think the amount of ancillary equipment to aggregate the bandwidth of the number of T1's required may make that cost prohibitive. I have not heard or read what the projected aggregate bandwidth requirement might be for an average cell tower. If they intend to allow the maximum bandwidth mentioned in the 802.16/WiMAX protocol, that would be a significant increase in the amount of aggregate bandwidth required at each cell tower. Maybe panTribe could comment on that. For example, let's say that 45Mbps or one T3 would be sufficient for a particular cell tower. That amount of bandwidth could be replaced with 30 T1's, but that would require 30 fiber/copper conversions to get the circuits from the fiber backbone to the cell tower. That also means there has to be that much copper in the ground running to the cell tower as well as sufficient demarc equipment at the cell tower. I believe that is why they are leaning toward running fiber to a centrally located cell tower and microwave backhauls from nearby (within range) cell towers. More than likely it will be handled on a per site basis with physical and RF environments dictating the most efficient and cost effective topology.

Michael Kassner
Michael Kassner

Please forgive me though as I do not understand your second comment. Sprint did or did not use your findings? Also if you could expand on what you found out, I for one would be very much appreciate your time and effort.

Sensor Guy
Sensor Guy

In this arena of technology, there are so many factors that can guide you to one vendor or another so it's very tough to suggest any one vendor. Tower design, terrain, distances, weather (particularly hurricanes), seismology, propagation, licensed spectrum, legal, jurisdiction, encryption (security), power, etc. are just a few of the factors to consider. Some vendors on the tower wireless like Motorola like to dictate who their bakhaul vendors should be as well. Some of the vendors I've encountered or used in this segment are California Microwave, Harris, Proteus (Microwave Networks). There is considerable chatter in this part of the business about trunked backbone networks versus meshed networks as well. Trunking is the traditional method and meshed is the rising star. Many folks are doing hybrid networks, particularly in cases where intense local wireless communications are needed for a short period of time. This is an area with little education. I teach an RFID class and thinking of expanding into this niche area of communications, because there's lots of interest and very few people with skills out there on how to roll out practical microwave and other backbone projects.

Michael Kassner
Michael Kassner

I do quite a bit of work on the rural/urban fringe, many clients would give up body parts for DSL. I have been reading about new technology that will extend its reach ands speed. I hope that happens soon as they figured out how to keep DSL as robust as a T1. Qwest is also giving business grade SLA's which increase their appeal.

Michael Kassner
Michael Kassner

I also am one of those "old engineers" (worked with tube transmitters, punched cards, and IBM "big iron") thus have very similar view points to yours. One aspect that I didn't consider is the different mindsets between data and telco types. My work bridges that gap on a daily basis, so I would have missed that analogy. But the very fact that all the US telcos are advertising for microwave engineers seems to point that out. Very interesting. I very much appreciate your insight and knowledge on this subject matter.

Sensor Guy
Sensor Guy

As an old engineer, I was taught about the priciples discussed in the "How to Cheat" web page in the classical physics and mathematics methods, not necessarily appropriate for the masses that today need to learn this because of electronics commoditization! I did find the RFID pages very interesting in that they use a very innovative approach to teaching the physics sorrounding the tag and reader interactions. RFID is not rocket science. It isn't even new science. It's an optimization and use of well known RF principles with the advent of more sensitive equipment. This technology has been in use for decades in railroads and other places and used to be known as auto-id technology. The real innovation is that the miniaturization and electronics survivability of RFID has created a "wild west" space for innovation in the application and use space, very similar to the early days of the Internet and then later the web. The clash between mesh and trunking is going to be very interesting. I think both technological paths have their faults and strengths, and the end result will be more of a hybrid, stronger environment. We must also remember that mesh comes from data thinking and trunking from more voice and video mindsets, so we are seeing here also the clash of application cultures. Should be interesting and the consumer will gain most from it, just like we've seen with VoIP. I view RFID as a technological life form early in its life. If you consider the PC a technological lifeform as well, then you know it spawned other branches in its "technological family" like laptops, etc. I expect RFID to move up in complexity and capability and well as horizontally in efficiency.

Michael Kassner
Michael Kassner

I find it interesting that they are looking at mesh networks for backbones. I work in 802.11 mainly and all of my enterprise clients are moving toward mesh networks, especially with 802.11n and multi-radio units. I also am getting more and more impressed with the mesh routing algorithms as they are getting mature enough to really work well. I have to give credit to the MIT "rooftop gang" for developing some stellar applications. You happened to mention RFID and that you teach a class about it. I just did a rather odd book review about one of the How to Cheat books names" Deploying and Securing RFID". http://blogs.techrepublic.com.com/networking/?p=475 I thought the authors did a good job. When I mentioned odd earlier it was because in the post I rambled on a bit about being impressed with the way they explained RF propagation and digital modulation. Truth be told, I always read anything that Frank Thorton writes, as he is one of my heros. I certainly would appreciate any comments you may have about the post, book and RFID.

Editor's Picks