Which Way For Zambia? (Wireless or Fiber)

ZAMTEL is currently constructing a fiber optic line to connect Zambia to Zimbabwe. What is interesting is the fact that ZAMTEL has in collaboration with ZESCO at a cost of between $11M and $31M found it viable to mount the fiber optic network on ZESCO’s electrical poles.  Fiber optic technology represents both the past and the future in relation to other new technologies like Wireless (WIFI) technology. I would in this article endeavor to draw parallels between Wireless technology and fiber optic technology. Bearing in mind that while we have electrical poles everywhere in Zambia, we also have a very solid wireless infrastructure in place upon which Zambia national broadcasting corporation TV transmissions operate.  We can virtually build Zambia into a first class WIFI country just by using the towers that are present virtually anywhere in the country.

The principal dimensions along which I find it most instructive to compare fiber-optic and wireless access technologies are bandwidth, initial implementation costs (installation) and recurring (operation and maintenance) costs.

There are pros and cons to both fiber optic and wireless technology in a broader sense. In this article I will limit my exploration to implementation based on Zambia bearing in mind costs, terrain and existing infrastructure.

IMPLEMENTATION COSTS

Deployment of such a magnitude would require substantial capital injection into the project. Wireless has a decisive advantage over fiber on cost, labor and maintenance. The latter incurs costs of preparing a physical path across the terrain between two points; the installation of cables along the route may entail expenses such as trenching, mounting poles etc. These activities can prove to be labor intensive.  A fiber optic cable in the US costs between $17.00 – $32.00 per two (2) meters cable, add transportation and VAT you are talking $50.00 per two (2) meters.  The cost for cables couplers and labor is staggering for fiber optic. Wireless on the other hand does not attract cabling and pole mounting or trenching costs. The technology utilizes the radio waves to transmit a signal from point A to point B.  Acquisition and installing the data-transmission equipment at either endpoint for fiber optic is another monumental cost comparing to what such equipment used for wireless would cost. In a fiber optic transmission you would need a fiber optic transmitter and a fiber optic receiver. For wireless depending on your implementation, you would need radios on both sides of the connection. Installing fiber in rural areas can be quite expensive, especially in the mountainous regions; rugged terrain presents obstacles along a direct path between two points. These installation costs can be a barrier to deployment in rural areas as they must be offset by customer demand in the short term. In contrast, the installation costs for wireless are limited to the cost of the base station(s) and Customer Premise Equipment (CPE) devices for end users. The advantage of wireless for rural deployments across rugged terrain is clear – no trenching, no pole attachments, no cable-installation costs between points A and B. As long as the signal can reach its destination without encountering obstructions or experiencing attenuation, it doesn’t matter what lies in between. Bearing in mind that in Zambia we have a microwave network across the nation with towers less than ten (10) miles apart. With the new wireless technology called WiMax broadband 802.16e you could connect two points about 20 – 30 kilometers apart.  This visibly present infrastructure presents an opportunity for ZAMTEL to create a wireless mesh across the nation. With proper vision and leadership Zambia can become a cyber nation. What we are spending on the fiber optic backbone deployment is enough to turn Zambia into a first class cyber nation. We should start thinking about linking directly to the satellite backhaul. The main link is in the extra terrestrial. The whole concept of linking Zambia and its SADC neighbors on earth should be substituted with a vision that puts Zambia on the path to the cyberspace. We should connect to Zimbabwe via satellite and not via land technologically. The techno meeting place is the cyber space. One thing to keep in mind, however, is that a common initial expense to both fiber and wireless is the backhaul connectivity to higher-tier networks (satellite providers). While this is not a factor in the comparative evaluation of fiber and wireless access technologies, (except to the extent that some portions of the “middle mile” may be done wirelessly at lower cost), it is an essential driver of total project cost.

OPERATIONAL AND MAINTENANCE COSTS

Beyond the initial installation cost, though, there are additional cost factors that are recurring in nature. For fiber, the cost of a dedicated path between the endpoints frequently carries with it re current maintenance which would range from replacing broken cables or in this case ZESCO the fiber optic network is inextricably bound to the many difficulties that ZESCO faces in maintaining its power line network. The costs of keeping a fiber connection operational generally exceed the costs of maintaining a wireless link between the same two points. The reasons for this, not surprisingly, are a result of the need to maintain a physical data path in the case of fiber. Fiber networks, because they employ physical cables, are vulnerable to such hazards as cuts from backhoes, downed poles resulting from storms or car accidents, and even squirrels chewing on the insulation. Once a fault occurs, it must then be located along a cable route potentially spanning several miles. On the other hand, with wireless, there is no physical infrastructure to maintain between provider and customer, and faults are by definition at one end or the other. Lighting is also another major factor in most tropical countries.  If we go by ZESCO’s maintenance routine you will realize that the fiber optic cables which will be mounted on ZESCO poles.  Will also be subject to the very failures that ZESCO experiences when a pole fails or a pole is struck by lightening.  The other factor I’m sure most of us don’t want to even mention is the fact that fiber optic cables are expensive which in itself makes them a target by thieves. ZAMTEL went through this phase in the 90s when their cables where being stolen every day. I’m sure they are still grappling with effects and costs associated with stolen communication lines. In Orange county California have of late experienced a 60% rise in thefts relating to copper telephone lines. This has been attributed to the high copper prices. Fiber optic is even more expensive. On a positive note it’s very difficult to sale fiber cables mostly because of the limited market and also the fact that not every fiber cable will work for every fiber optic network.

BANDWIDTH

 First, in terms of bandwidth, fiber stands head and shoulders above wireless. This is probably the most frequently cited distinction between the two. With fiber, you get a dedicated data path between two points, and the bandwidth of that data path is limited only by the capabilities of the equipment at either end. Theoretically, there is no limit to the bandwidth of a fiber-optic connection. With wireless, on the other hand, you’re not using a dedicated physical data path; you’re using an RF transmitter and (perhaps multiple) receivers to turn the air between two points as the data path. And while the ability to set up point-to-multipoint communication is a decisive advantage that wireless has over fiber, the bandwidth of a wireless signal is constrained by a number of variables, including the amount of spectrum (the number of frequencies) you are authorized to transmit/receive on), the frequency itself (how many cycles per second you can transmit/receive), and your modulation scheme (how many bits you can push over one cycle). If you’re working in the unlicensed spectrum (which a majority of wireless broadband hardware providers use in their products), the number of frequencies you can use
to push data between sites is quite limited, relative to some of the larger spectrum chunks the CA offers for licensing to both ISPs and Telco’s. There is a trade-off between frequency and data-carrying capacity, such that as you lower your frequency (which gives you an increased ability to penetrate trees and buildings – a highly desirable characteristic)
you lose total bandwidth. But this is not a hard and fast rule, since the other relevant variable here is the modulation scheme implemented in the wireless hardware. Newer generations of base stations and customer premise equipment using
new current standards, as well as emerging wireless communications standards (802.16), utilize improved modulation techniques to squeeze more bandwidth out of the same frequency. But at the end of the day, the bandwidths achieved by
wireless technologies are still orders of magnitude behind what is possible with fiber. Where most wireless setups can deliver bandwidths of multiple megabits per second, the most advanced fiber-optic connections are delivering multiple gigabits per second. The suitability of a connection for a particular application depends not only on its downstream bandwidth in Mbps but also on its upstream bandwidth, and the extent to which it can be configured to provide Quality of Service (QoS) – guaranteed reliability. Whether or not a fiber-optic connection is symmetrical to the end-user will depend on the layer 2 protocol (e.g., Ethernet, ATM, SONET, PON) it is using. The Ethernet protocol, which is what yields “gigabit” transfer rates and is increasingly being deployed over long and longer distances, is symmetrical. Wireless offers flexibility in upstream/downstream bandwidth provisioning, but QoS more difficult to guarantee if bandwidth is shared between users. As a general rule, one can expect better QoS from a fiber-optic connection, as it is a dedicated link between two points.

SCALABILITY

Scalability is a very intricate subject considering how fast technology change. In this can we will look at relative scalability of fiber and wireless, which also addresses their relative longevity. There are two different dimensions of scalability relevant here; the first is scalability in terms of bandwidth (the ability to add more bandwidth in the future), and the second is scalability in terms of architecture (the ability to connect additional sites in the future). In terms of bandwidth scalability, fiber systems are more scalable because the same infrastructure can support future advances in data transmission. One need only upgrade the endpoint equipment to often dramatically increase the performance of a sunk investment in a single cable. Fiber-optic cabling is effectively “future-proof” in this sense. Wireless systems typically support only a single frequency/modulation scheme, and the endpoint equipment must be completely replaced to take advantages of different frequencies or modulation schemes that can increase bandwidth. The important point to note here is that for both fiber optic and wireless access technologies, you need to upgrade the backbone to increase bandwidth.  (The ability to add more bandwidth in the future), When we look at adding new sites to the network I will use an example of a car versus the helicopter. Fiber optic is like a car driving to kabwe from Lusaka the only optional routes you have a pre determined by the road construction planners. If for instance you decide you want to drive to Nampundwe you have to go to kabwe first then drive to Nampundwe or you have to construct another 70KM road from Lusaka to Nampudwe at a very high cost. On the other hand with wireless (helicopter) all you need is a soccer (football) field to land. If you analyze the impediments that come with a fiber optic network, one could make the observation that although we have heard many; many times that “fiber is future-proof!” really there isn’t that much difference in the bandwidth scalability of the different access technologies. With fiber, the future-proof part is the transmission medium – the glass optical fibers that run between sites. The idea is that once you lay a cable, the performance of that same cable can be increased by upgrading the endpoint equipment. With wireless, on the other hand, the medium is the air itself between the two endpoints, and the amount of data you can transmit over the air can be increased by… you guessed it… upgrading the endpoint equipment. So when we are careful to distinguish between the medium of transmission (glass fibers vs. air) and the electronics that push data across the medium of transmission, we realize that there isn’t *really* that much distinction at all between fiber and wireless in terms of bandwidth scalability. In the end, any advantages one has over the other is going to depend on the relative costs of fiber-optic endpoint electronics  and the cost of mounting the network to poles and the huge labor intensive maintenance vs. wireless transceivers. The other dimension of scalability concerns the architecture – the ease of adding an additional site. Here there really is a difference between fiber and wireless, because as was mentioned before, initial planning and getting it right the first time is critical with fiber, whereas wireless networks can be expanded much more easily in the future, as long as careful attention is paid to the load on the base stations and the backhaul connections so as not to degrade the bandwidth quality for the previously existing subscribers.

Conclusion

There are advantages and disadvantages to both technologies. Considering the costs involved in implementing the fiber optic network and the unstable ZESCO network. One would always gravitate towards Wireless which is far less expensive to implement and maintain.  The question of flexibility in terms of upgrades and adapting to the constant change and improving information technology also emphasizes the fact that wireless is the way to the future.

Francis Malama
Los Angeles, California
francis@zambiahosting.com