Running a WISP takes a wide variety of skill sets. My collection of technical skills looks sort of like this:
- RF Engineer
- Field Installer
- Network Administrator
- Client and Desktop Support
Besides being able to crimp a cat-6 connector, other skill sets needed include:
- Shipping Clerk
- Purchasing Manager
- Marriage Counselor
Most small WISPs can’t usually afford this size staff, especially starting up. Fortunately there are ways to minimize the amount of training and cost for the technical skill sets. Keep in mind that nobody works 24 hours a day although to many of us it seems like that. Redundancy however, in infrastructure and personnel, is critical. If you don’t have 2 people with multiple skill sets to cover more than one area, the best thing to do is see if a skill set can be eliminated or at least the level of experience necessary to support it is reduced.Either that or figure out how to clone yourself.
You can always contract the technical skill set you may be short on. The good thing is that you aren’t paying for a full time salary, you aren’t paying for downtime, and you don’t have to worry about taxes or health care costs which because of our government, are a bigger variable than our national debt. At the same time,these people aren’t available 24×7 unless they also happen to be your brother-in-law or you have pictures of them they don’t want made public from your college dorm days. In addition, if you have to use them a lot, they may cost more than a full-time employee.
WISPs are usually built around the technical skill set of the founder or founders of the company. Mostly this is either networking or RF experience. Small WISPs are usually created because there is an area that needs internet service and those entrepreneurs may simply have basic computer skills or less. Fortunately, it’s still possible to get a core system running with a little outside help and to keep the operational costs very low.
I’m a huge believer in the stitch-in-time-saves-nine adage. Keeping that in mind, planning for that lack of resource can be the difference between a startup making it or failing. The fact of small businesses, especially startups, is that the owner/operator is the last in line in payroll and that employees always come first. Plan for this before you jump in and find out your 5-year old car is going to be your 10-year old car before you can afford to start taking a salary to replace it.
So let’s start with the RF Engineer. Designing the network is only done once so that’s an easy one. However, if you can’t run Radio Mobile or something similar and aren’t comfortable in calculating path loss, then it’s the one place where you need to hire an expert. Every day, even more tools keep coming out or become easier to use. For smaller WISP areas, most designs only take a few hours to lay out the RF portion if it doesn’t need on-site analysis. Future expansion can typically be extracted from the core design which means little futurecosts unless there is a big change in the design. Score one for contracting on this one. Yes, full disclosure is that this is what I get paid to do generally but you won’t catch me fixing a clogged garbage disposal since I am a big believer that 10 fingers is more functional than 9.
Next on the list is the Network Administrator. Contrary to popular belief, it’s possible to build a network with a basic understanding of networking and no command line experience. Smaller WISPs of a few hundred users or less really don’t need routers at every tower. Clients don’t care how elegant your network design is. If you don’t believe this, put in your sales brochures how you really efficient the network is based lack of broadcast traffic due to the number of VLANs that are implemented. Since I have never seen that technique used in a Madison Avenue advertising campaign, I’m guessing it’s not the best move. It’s also a bad starter conversation on a first date unless you think the hypnotizing your partner into a catatonic state is your goal.
The easiest way to make this work is to simply make sure that ever CPE device (assuming an outdoor CPE model) is in router mode with its own DHCP server and private internal subnet. You can even break the network down further by using one subnet on half your CPE’s and a different subnet on the other half. This solves 2 issues; it allows you to create a pseudo VLAN structure simply by limiting the number of users on a subnet on the external interface and by not allowing other users to browse the network. In reality, this design could take you to several hundred users.
One of the reason this works is because of imbalance. Assuming each AP can support 30Mbps and you have 8 APs running. The total internal capacity of the network is 240Mbps. If the internet bandwidth is less than the total capacity of any single AP, then a little extra broadcast traffic around the network really isn’t a big problem. In fact, the best thing to do would be to spend a little extra money on the only router you need to buy. Look into getting a router with a firewall. If you can budget for web-content filtering, bandwidth throttling, VPN tunneling, etc…, that’s even better. Look at SonicWall, Watchguard, Barracuda, Zyxel, etc… for products that give you extra featuresin an all-in-one Web GUI type of solution. Keep in mind that you still need a backup device for redundancy so keep a cheap basic router capable of handling the capacity to swap in when needed.
Next on the list is the Field Installer. If you have enough clients to keep someone busy 40 hours a week doing installations, you really wanta full-time employee to manage quality control and keep the efficiency high. It’s much cheaper to train one person than several. It’s also much better for the clients to see continuityof staff. Since some areas may use several different techniques for installations, it takes quite a bit of training and experience to improve the efficiency and that doesn’t happen when you use part-time and a lot of different installers.
Last on the list is client and desktop support. There are many different ways to handle this based on scalability of operation and different levels of support. At minimum, a WISP has a vested interest in making sure the client can connect to the system. That means yes, you actually have to dialog with clients when theycan’t connect. The smaller WISPs typically handle this in house. If you have a solid network, then the numbers of calls should just be a handful. Using automated monitoring tools like TheDude, AirControl, and others along with VNC on client computers when they do have a problem is simple to deploy and inexpensive. I’m also changing over all my indoor routers away from the client buying whatever they want to an indoor wireless router we supply and manage.
The bottom line is that it’s possible to build a basic network up to several hundred users and keep costs down. It’s more important to focus resources into areas that directly affect profitability and viability, especially if you have limited funding and staff. This design also makes it possible for even basic computer people to support the network. So keep pinching those pennies and let’s get these WISP businesses up and running.