My memory must be slipping. Dialup backup (via PacWest), subsidiary services such as email, web space, web hosting, domain registry, credit card laundry, site builder tools, CGI/PHP/Perl/and such, VoIP gateway to PSTN, inventory of installation hardware/radios/antennas/consumeables, method of qualifying customers for line of sight, chain saw for trimming the inevitable tree in the way, channel plan, bandwidth management (very important), authentication server (RADIUS), spare everything, UPS, solar, generator, ad nausium.
Sorry. I didn't mean to criticize anything you posted. I just wanted to append a more realistic and more detailed "things to burn money upon" list for starting WISP. Please note that I didn't include office space (home office?), office supplies, tools, installer supplies (bucket truck), tranquilizers, anti-depressants, and storage for the inevitable tangle of coax, antennas, and "things to be repaired eventually". A large trash bin or dumpster is probably the most important tool.
I have more than a little experience trying to throw together a WISP division with an existing ISP. I gave up. The biggest mistake is to assume that traffic and support patterns will be the same as with a wired ISP. They're not. Without bandwidth management, it is fairly easy for a single customer to monopolize all the ISP's available bandwidth. However, bandwidth management at the ISP end only solves half the problem. Because airtime is a precious resource, proper bandwidth management requires that the client radios be prevented from transmitting, rather than just dropping packets at the ISP router. That means that the WISP is now in the remote management business of the clients radio (bridge). That's usually no problem with hetrogenous installations, where all the hardware is the same. However, wireless hardware obsoletes itself very few months insuring that every client radio and installation will be different. At best, there will be incompatible firmware versions. At worst, new radios will require new management tools (or methods).
Support is another problem. Wireless customers generate about 4-10 times the number of support calls as wired (lease line, dialup, DSL, cable modem) customers. It's just the nature of the beast as wireless is much less reliable. Without proper diagnostics, the customer has no clue as to what is happening. All they know is that their link to the internet is down or constipated. From where they sit, they can't tell if it was caused by the neighbors leaky microwave oven, bird sitting on the dish feed, water incursion, co-channel competitor, new cordless phone, springtime tree leaves, or their own wireless TIVO, that is causing the outage. So, they call the ISP, who might have even less information. I did some work in remote RF imparement detection and analysis, which ended up looking like an ELINT or ECM system and far too expensive. Trying to troubleshoot link imparement, when there's no connectivity from the ISP end is not a trivial exercise. I tried wireless routers with a built in modem "backup" that also allowed me login to the customers end via telephone.
The last straw for me was when the customer ordered a pizza dish for DBS TV reception and the installer "borrowed" the WISP radio mounting pipe for the dish. Since the antenna was internal in the PVC NEMA box, he had no idea which way the WISP antenna was pointed. So, he managed to plant his pizza dish directly in the link path. Despite me repeatedly asking this apparently clueful customer if anything had changed at his location, it still took a 45 minute drive to his location to discover what should have been obvious. Then, the customer was wondering why I didn't want to climb his 3 story high roof, in the dark, in the rain, and without proper safety equipment. Since he was not directly involved in causing the problem, he refused to pay for either relocating the pizza dish or the WISP radio. This was NOT an isolated case, but very typical. WISP's are expected to provide 24x7 service, just like wired service, and restore service for free, regardless of the cause of failure. Since one unpaid service call negated about 14 months of per customer margin, this was not a supportable business model. So, I wiggled out of the biz.
The other big problem is scaleability. What works for a small number of clients falls flat on its face when faced with a large number. I don't wanna go into much detail, but lets just say that 802.11 is gross and inefficient for large numbers of active clients. A polling scheme, as found in Karlnet and WiMax is much more efficient. The airtime efficiency problem doesn't show up until the system has grown sufficiently large to guarantee that protocol changes will be disruptive and uneconomical, especially with customer owned CPE. Anyone can make a small system work. Larger systems, required to support a realistic WISP business model, usually require alternative protocols.
I've just finished a 12 month contract with a Wireless ISP (Pipex in the UK). They had their own wireless network in the 4.8Ghz region. My aerial had a powered cable connecting to an ethernet box. The ISP's end was about 10miles away with line of sight. Although nominally a 512/512k service I rarerly got better that 200k but still a lot better than 56k dial-up.
The term they used was WDSL and it was not easy to find that they offered the service.
Fortunately our local exchange was recently ADSL enabled and even though the max is also 512k it is twice as quick as the WDSL.
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.