Wireless Wonderful!
----------------------------------------
***ESSAY DISCLAIMER and ADVERTISEMENT***
----------------------------------------
As essay to promote the thoughts and results of some testing and considering by Stephen Hill, head of Zo.com, and having contracted under the name of 'ArcTypics' since 1990. I, Stephen Hill, make no presumptions to the effect that I have performed testing of laboratory quality, nor that I have given even comparison to all products. In particular, there may be several products, solutions, examples, situations, individuals, and just outright pure idea that I may have missed in the process that lead to the creation of this essay and its results. It is plausible that I have made statements and suggestions for products or solutions that are based on prejudices I maintain, or conservative portions of my brain that constantly fight my personal attempts to grow and accept new things. If I recognize these situations in my writing, I will attempt to specify this. If I recognize parts of these writings that are more accurate or well tested than this disclaimer implies, I will also attempt to specify that.
But the biggest disclaimer of all I must make is that I have grown up with the 'internet' industry, I would classify my job fuction and abilities short as 'automation and efficiency of procedure, design, and layout.' I can come to your group, your company, your home, I can sit and chat with you, find how you do what you do, find what your dreams are, what your employees desires are, and we can make it happen.
I have spent the last 10 years working alone, with individuals or smal companies, while maintaining career positions in varying internet industries. I think I've reached a point where I can combine the experts I know who have found no home in the world for their careers, with the real world people I know who are ready to work to make a difference instead of being another small spot on an org chart at a large company, to create a family of people who are here make your dreams work, to make your services work, to make your company compete.
This essay, and several others like it that I will be writing are here to help you. To sell you on zo.com, and to get YOU to respond to ME so that I may know more, because sometimes browsing the web and reading books can't tell you everything. If you want to hire me or my network management or colo services, email life@zo.com. If you want to know me personally, email nehpets@zo.com, if you want to respond to this message, email wireless-list@zo.com.
--------------------
***END DISCLAIMER***
--------------------
Networking layout and implementation for Example home/small office network.
Some extensions of this can be implemented with modification on a large scale office network, while maintaining the low cost per unit design, and keeping the simplicity of maintenence by small unit design as well.
--------------------------
*Hardware for connections:
--------------------------
1 workgroup laserprinter ethernet builtin no wireless option (?)
1 personal laserprinter ethernet builtin no wireless option (?)
2 special [usb] printers usb builtin no wireless option (?)
1 tower server box ethernet card wireless pc card adaptor option
1 tower desktop box ethernet card wireless pc card adaptor option
2 laptop computers ethernet builtin/card wireless pc card option
2 future/visiting unknown ethernet builtin/card wireless pc card option or adaptor
1 incoming wireless/wired ethernet/usb/pcard no control over network provisioning
-------
*notes:
-------
No bulk purchase discounts located to date on any quality products. testing with SMC, netgear, linksys, dlink have shown netgear with highest degree of simplicity, quickness company support, lowest chance of DOA, and highest future compatibility rate. (Note that 'lucent' 'orinoco' products were continuously highest rated by the industry, but the company attempted to create a seperate organization to sell this product line, and in the procedure, all of these products appear to be disappearing from the market.
Additional 'speed' notes for comparison of 802.11b '11meg' via shared basestation to wired '10meg' via shared hub, wired '10meg' via switched, and 100m shared and 100m switched:
No statistical data is available, nor were continuous repetitive test run [noone was willing to pay me to waste that kind of time]. However, at 4-8 machines on 802.11b wireless at 11m share, with approx 20% crosstalk, and the remainder being 'outgoing' to the internet via the basestations 10meg uplink, I found crosstalk communications generally faster than via share 10mg, similar to switched 10meg, and only moderately slower than 100m switched. [actually faster than 100m shared, go figure]. internet traffic concerns were unaffected in any situation, as the uplink connection was limited to 10meg on the switch internet provider port, and the outgoing traffic effectively maximized the port capabilities. My suggestion for this would be a 100m uplink capable basestation in high-traffic workgroup situations with high speed network connections, or a failure to care and patience in situations where the internet connection is less than 10m in the first place [such as the iec/rocair home network]
'bluetooth' is a non-standard modification to 802.11b which reduces power when communication sources are located closer together, unnecessary, expensive, and potentially problematic in future solutions. (Althogh its great for laptops, pontentially PDAs and should hopefully work with the same or slightly upgraded basestation!]
---------------------------------------
*Wireless costs and Wired vs. Wireless:
---------------------------------------
Cost of wired [card/pc card, plus wiring, plus switches/etc, plus upgrades and size/etc for smaller or laptop units] makes this a desirable solution for ALL products, not just laptop or similar solutions. One mid-quality wireless basestation with 64bit or 128bit encrytion utilization can handle 15-45 units without difficulty at a full 11m. My suggestion is as follows:
*Laptops*
All laptops be installed with wireless 802.11b ['WiFi' compatible] products, netgear preferably [$150 for 128 capable encryption card]. This leaves a minor 'antennae' portion extending out from unit. future laptop purchase should involve buying built-in/internal WiFi capatbility to eliminate this minor inconvenience. Average cost per machine for hardware and labor costs in teh built install would be approx $250 [this includes my fees for buying, installing, testing, and convincing the box that is the primary network point].
*Desktops*
All desktops should be installed with pci card adaptors which use the plug in of the laptop style pc cards. This adds approx $30 to $60 to the price for the card itself, plus the cost for the laptop pc card. This allows replacement of wired network card slot, maintenance of 100% same hardware/services support, and at any point of upgrade in the future, ALL units may be flash or hardware upgraded to maintain the exact same components/features. Cost per desktop is laptop cost, plus $60 for the adapter card, plus $50 for extra labor involved in physically changing the card out. Totalling $310 for the machine changover.
*Basestations performance*
For the basestation, the options become more difficult and also much less plausible to test. I have limited my detailed testing to utilizing the netgear basestation offering. This has one WiFi capable input [up to 45-60 unit for communication if the network communication is limited, expect up to 15-20 unit at high communication levels. This basestation handles 64bit encryption maximum at this time [which I consider highly acceptable. Wasting time, network speed and resources performing excessive encryption for data who's interception and 'hacking' requireing completing simple and alternate efforts [discussed below] is not necessary. If, at some point in time, a need for 128b or HIGHER becomes apparently, upgrading of the simple, low cose, single unit basestation is an easy way to affect this upgrade, and upgrading beyond 128b may require flashing all machine units or changing all hardware at that point, which nullifies any cost savings in planning ahead on this machine. I am happy with my netgear choice for maintaining a single point of service support and purchasing simplicity. It is also limited to 10m on the ethernet port side, but any solution this is placed into, again, this should be fine unless you're in a large network implemention. Thus, I will make the suggestion in this category in the next paragraph.
*Network utilization*
For small networks [10-30 people] with moderate simultaneous usage [shift working nocs or multiple machines per person without running servers] causing no more than 3-10 machines to be active at a time, allowing for occasionally off-wireless speed transfer delays with large data downloads or similar, I suggest staying with the netgear offered wireless basestation. It does not offer the best or most mini features or 'coolness' factors, but if a call or need ot communicate with netgear produces a conversation which allows you to specify you have an entire network of netgear pc cards, pci to pccard adaptors, and netgear basestation [as well as a few minor things I'll mention below], I can assure you that someone at the management level will then provide you with service well above the par expected because they should realize they risk a very large loss of word of mouth marketing by not completing your service need with perfection.
For larger networks or higher traffic to the internet [sitting on top of oc3 or higher internet connections, or at a colocation environment, or larger physical space situations, I suggest a couple of alternate solutions, which I'll address after discussing the non-standard devices and backbone network solutions for all of these solutions.
Non-standard devices:
---------------------
In any situation, the example in particular, many devices are not capable of being moved onto a wireless network directly. Either they have no internal provisioning for pccards or pci cards to convert to pc cards. Traditionally for any network upgrading on these machines [serial to phonenet, usb to ethernet, etc] have involved expensive 'media translation' devices. In this situation, you're left with these standard situations, or a new, unique, alternate but without any cost reduction in and of itself.
In my history of dealing with networks and machinery, these kinds of 'media adaptors' [phonenet to ethernet, usb to network, etc] have traditionally cost in the $300 range. Basically, someone creates a hardware bridge [device to combine two networks with different media types, as in phonenet and ethernet, etc], repackages it with a different name, sometimes a different look, and occasionally some kind of minor intelligence [a 'network spooling devices' to let your usb/serial printer become a network printer on the ethernet side.]
The solutions for these situations are always highly dependant on YOUR setup. Lets run through a few simple solutions that I think allow for most people situations, a combination may be required to fit your personal needs.
------------
) Often USB or serial style printers [if they still exist] are in usage by a single machine,or avoid
ethernet/network options because those are serious cost increases [however cheap they may seem to you, people looking to buy a $100 color printer think an extra $10-15 is a LOT and will skip that product over]. These can become network printers by buying $300 devices that turn their serial/usb interfaces into ethernet. Or perhaps more efficient and just provide a speciality 'minicomputer' with multiple usb/serial ports to let all your printers plug in, and ethernet, or a pc-card to hook up wireless networking. So you can pay $300-$1,000 [or more] to take a group of non-network or similar devices, and get them migrated into your new or existing network, with added print spooling functions if they weren't already there.
Sometimes people do this themselves, they buy a dedicated server machine, or use their personal machine to plug all these serial/usb/parallel devices into the ONE machine, then use network sharing to allow all other users to print through their machine. This can be free if you've already got the machine, or cost $300-$1,500 to buy a nice server machine to handle it. It can be mac, windows, linux, sun, freebsd, or a custom device you don't know or care what it runs, it just works. But it requires all your devices to be either physically near this service machine, OR have long cables running to wherever they may be.
Solution for usb/serial/parallel products? Well, I try not to be clouded by personal prejudices, but in this case, I'll allow myself some leeway:
If you bought it as a personal device, for yourself, your machine, you're fine anyhow. It's got its little physical home established on your desk/cube/room. Its connected to your machine, it works. If you share it on your network, thats fine, it doesn't care if you're wired or wireless. Relax. It doesn't need any upgrading, it doesn't needs it's own private wireless link. Sure, you can buy custom serial/parallel to wireless boxes, but who needs 'em. You bought cheap for a reason. If you WANTED super network color shared power printing, you would've and SHOULD'VE bought a cheap color ethernet laserprinter. If you STILL want that, same money, sell your color inkjet or little cheapo printer on ebay, buy yourself a cheap network laserprinter. It's not magic, its common sense. If you've all got these little personal things, or you're combining work networks or whatever, hook the best 2 or 3 up to a server machine [DEDICATED machine, don't use some high-use personal machine, take a machine people you let friends print their thesis from, or a share office machine for visiting managers or VPs, call it a 'server' instead of a 'spare'.] Make that machine wireless, give it your usb, parallel, serial printing devices, one of each type, sell the others on ebay or give them to less fortunate friends or offices. Then you can focus on upgrading one printer for everyone, or sell 10 stupid little color inkjets and buy one nice network color laserprinter. Scanners work the same now. There are ways to put your scanner on this one machine [you can hook up tons of USB devices to this one box, just dont use them all at the same exact time, or buy a high-quality usb pci card]. Sell your personal scanners, get one or two good ones. Create a 'social center' get a couch, a water cooler, a small fridge. Now your office people have a place to relax and a reason to be there. 'Oh, I'm just waiting for my color printout, or my chance to scan'. If your managers, your company complains, tell them to buy you better, faster equipment. [if it was a color LASERWRITER, you don't have to wait. If it was a $300 color high-speed scanner instead of a $50 fry's special, we'd be done].
And now, maybe, just maybe, you'll know your coworkers names. Or get that date with a coworker all the psychologists tell you not to get. Oh yeah, and since your entire office is wireless, if your IT read my whole essay, you might have a laptop, and you can get work done sitting on that couch. [read my other essay, entitled 'What work place?' to be released almost simultaneously.].
------------
) Network capable devices [ie, ethernet, or some rare hardware with pc cards or pci card infrastructures].
These devices present the biggest problem in the wireless ideal. They were purchased to provide self-contained network spooling, yet be presentable as speed and convenience to all LAN and possible WAN devices. If you upgrade to wireless services, they both need to be available, and need to be evaluated as 'second-quality' network citizens in traditional thinking. If they're already wired in, you're NOT likely to move a printer around, so why should you bother going wireless with them. Some wires will still be around [incoming network connection for instance], so why not just leave our printers on them. A backbone ethernet to a small switch [if we're staying brand loyal, a little 4, 5, 8, etc port 10/100 netgear switch [not a hub with 10/100 switching per port, but a 10/100 port SWITCH]. If, however, the printers are *also* spread out amongst many locations, then this makes pointless our efforts towards fully wireless simplicity and zero infrastructure alterations/installs. We now have a couple ways to integrate them wirelessly.
*) These devices can utilize low-end wireless basestations as ethernet to 802.11b conversion devices. This allows the ethernet from the printer [or multiple devices off a switch if they're physically directly near each other] to be connected to the basestation, which links into the wireless network system. This has the added advantage of increasing the redundancy and highspeed guaruntees of the entire system [see expanding the range below]. If there are many individual systems spread in many different locations, this can become an expensive solution. If there are few central locations maintaining said shared devices, this is not only moderately cost effective, but network enhancing [again, see expanding the range below].
*) Some of these devices may themselves have expansion abilities already built into them that allow them to be upgraded to newer technologies, including wireless. Whether they be pc-card networking options, pci-card, or mini-pci-card slots, or even USB ports for USB wireless solutions [not discussed yet]. IN these cases, treat them like a standard base machine/laptop, and you integrate them at normal cost levels.
-------------
*USB devices:
-------------
I've experimented lightly with USB based network equipment. The problems that I've encountered to date are that when used as the only USB device in a chain, they perform adequately, and seem to do the job. But once you start adding a larger number of USB devices into this mix, USB shows that it's capable of insuring continued shared response to all items, but fails to provide any individual item with its preferred extreme performance level. More experimentation is definitely required to satisfy me in the category of high-speed USB devices, but at this time, from my few experiences, I degrade my suggestions to emergency situations only. It may be convenient to have a USB 802.11b device in a drawer/closet to utilize on a visiting machine or speciality device [playstation2?] that has USB support, but unreliable or pre-filled alternate ports [no pc-cards, or an ultra-slim notebook that uses the pc-card for its cd drive, or any other number of solutions]. But when planning for a specific machine, design it to utilize the pci+pccard combo, or a pccard for this solution. (Many newer laptops give you the option for a 'mini-pci internal wireless ethernet', my suggestion on these is actually to NOT take that option,, and get something with internal standard ethernet instead [for use at remote sights without wireless options]. This way your laptop can share the same client wireless technologies as your entire office, with the same upgrade paths, and potentially the same future improvements guarunteed. Leave the USB for your mouse, keyboard, sound, webcams, and inkjet printers.
-------------------------
*Overall Network Designs:
-------------------------
In the world we're making where wireless is the key, you can take a little more liberty with your design of the office environment. You'll still want to put some printers, scanners, incoming network connections, and maybe a dedicated server [or spare/guest machine] sitting in one common area with easy access and room to have several people lounging or loitering around as they wait for their service to complete. This small area uses a very small wired segment, a conveniently small ethernet switch with several short ethernet wires to satisfy these office components without modular connectivity.
The machines on your network now have freedom of mobility that translates into anything you want. If you're talking laptops, you can have your office design be a living room. Some ergonomically arranged coffee tables or lap trays [as if you were eating your dinner], people's backs to the wall so they have the screen privacy so many people find more comfortable, but their faces towards each other. Even our desktop machines have a bit more freedom of movement. Until now, it was a large box, tethered to its ground in location by so many wires and spaghetti, we were afraid to touch it long enough to install memory. Today, our network wiring has disappeared. You can grab some logitech RF wireless mouse and keyboard combination, and with a good combination of velcro, cheap and aestethicly pleasing USB hubs arranged artfully onto the exterior of the desktop case, the only thing your modern art is lacking is the ability to unplug its unsightly and thick power cable. Its leash, the dog chain that not only pulls it back into its knee-knocking home, but without any cute little wire gadgets so common in the rest of the cables we add to our machine, pull the machine too hard and the loud click and slow whine of the power loss because the cable slid right out of place we've forced ourself into a time loop delay of waiting till we are alive again, till our desktop shows that life shares our space.
That pesky desktop offers so many more powerful toys for us, but each toy tethers yet another leash. Our sound is realistic 3-dimensional with enough bass to shatter my mother's oldsmobile, but requires approximately 15 wires in a random, almost pentagram-shaped design. The monitors video cable thick enough to tow a car with, yet never reaching more than the 3 feet straight down to my machine. The camera, so artfully integrated into my laptop that I sometimes forget its watching me, now offers me 7 ways to hide itself amongst my monitor and speakers, none of which appear to work.
But I digress, I find a flat monitor that fits into any space I desire, with cable enough to run my machine from the back porch, and that same cable plugs into the speakers and usb and power ports on my desktop machine. Now I have a big super-cable, and growing out of my skinny monitor area speakers that I need a microscope to actually see, but sound like I've been surrounded by monks speaking in unison, and there's a camera in there that is invisible as well. And I realize that someone took the top half of my laptop, and let me plug it into my desktop machine. If I want to run some friends special hardware really quickly, there are cute little innocuous USB ports on the monitor so that these things can plug in where I'm sitting. My desktop machine is hiding is some corner I don't even know about, invisible because all its got it some power cable, and this magical wondercable snaking out to the flat monitor hanging from the ceiling, floating in space while my keyboard and mouse [optical, yet wireless, working *near* any surface, including my carpet], magically convert my fingers desires into things that zoom through the air and appear on my genie monitor floating in mid-air. Ok, I'm trying to wax poetic about rare and unusually expensive desktop attempts to become convenient. You still pretty much want a nice laptop. But we can make desktops that give you some range and fun. Kind of like the laptop monitor/keyboard/mouse of an ultra-light laptop, but with a really powerful hardware machine hidden in some corner.
---------------------------------
*Let's get real, let's talk cash!
---------------------------------
Ok, this will be a quote, based upon the office listed at the beginning of the message:
(prices include approximate shipping and taxes)
Item Hardware Cost Labor Excess Total
---- -------- ---- ----- ------ -----
*-wireless
Tower box PCI+PCcard 250 125 0 375 * 2 750
Laptop PCcard 175 75 0 250 * 2 500
Network printer Basestation 350 50 0 400 * 2 800
Router Basestation 350 50 0 400 * 1 400
Router Router 500 100 0 650 * 1 600
Switch Switch 100 0 0 100 * 1 100
Spare PCI+PCcard 250 0 0 250 * 1 250
USB alternate USB 200 0 0 200 * 1 200
Planning & Layout & testing 0 400 per hour 100 * 4hr 400
--- --- ---
2200 800 $4000
Setup:
1 incoming net connection [sprint wireless broadband] connects to outside port on router.
1 office side connection connects from inside port on router to wireless basestation.
4 machines [2 desktops, 2 laptops] utilize wireless networking to communication with each other, and with the internet via the basestation to the router.
2 network printers connect to the ethernet port on their own basestations to communication with the wireless network. In addition, if setup properly and the hardware works as advertised, the 2 extra base stations [located in different portions of the office/house] expand the range of wireless connectivity, and insure the maintenance of the highest connection rate wirelessly in as large of an area as possible. In the future, these alternate base stations can be used to hook up any other ethernet enabled device, or [via switches/hubs] expand those locales to include more non-wireless capable equipment. Effectively creating a 3-zone system. Alternate zones or work areas can be created by adding more basestations which require nothing more than a power plug available in an area. [ie, kitchen zone to include a back porch/living room/garage if the 'upstairs office' is not adequate, or 'downstairs colocation floor' if and 'upstairs office' doesn't have the power to reach.
All minor USB/serial devices connected to individual tower machines with sharing services enabled to allow other users to reach them.
One spare pci+pccard combo would work for either an additional laptop or desktop. Since the pccard would work for the laptop, and the pci adaptor would allow a desktop to use the pccard. This covers the inclusion of any one kind of machine immediately without further purchase or planning. One USB alternate hardware option performs a similar function. Allows for testing of USB capability speed and reliablity wise, plus can be used for nearly any form of device that can be networked. From computer equipment to consumer level equipment [Playstation 2s, some PDAs, etc].
The inclusion of a switch between the router and the primary basestation allows for the option of proving legacy support for hardware or situations that may need a solution with wired capabilities.
Under this solution, there would end up being a large quantity of spare cabling, network cards, hubs, etc. This equipment is hard to evaluate value of roughly without itemization. But for the purposes of this quote, I would specify that I'm willing to provide a credit of $250 in return for the bulk returned equipment. With variation if any particular items turn out to be of extreme value or extreme devalue. This quote for IEC/Rocair home office network would involve going pure wireless with a router and firewall capabilities of professional quality, and security at the wireless level that would be satisfactory. It allows for convenience of arrangement, and upgrading of technology primarily via software upgrades, and an ease of hardware upgrade at approx $150 per unit [$1,500 overall presumeably] should new levels of quality come available [such at 100m wireless or 50m or similar]. These solutions would also allow the laptops configured within these systems to automatically join into WiFi networks at many alternate locations as are becoming more common. And for approx $500 to $750 per laptop, they can be designed with a series of hardware that would provide network connectivity in perhaps 90% of the bay area without user effort to determine choice of connection.
This rollout can be staged to be performed in specified amounts/chunks, but the costs increase if done in smaller quantities for upgrades. The suggestion I have would be to make the complete changeover As soon as possible to take advantage of some lower pricing schemes for soem of the higher quality products attempting to break into the market vs. low quality devices, and to obtain 2nd generation equipment that, in one to two years, it may be possible to convince the companies to replace the physical hardware with 4th generation equipment that is software upgradeable for enhancements due to the degree of early adoption and qty of product in early adoption.
In addition, if this rollout for your office is accepted and completed, The contacts with the company who's products are utilized may produce future bulk discounts [which may be passed along retroactively to you], and the cost of my labour would reduce in actuality due to experience, though the charge would maintain the same. And following a successful proof-of-concept on your network, any follow-up installed via word of mouth or directly contacted via you would provide a 10% finder's fee from my personal labor rate return. [which would increase to $1,000 probably, with $100 fee for the word of mouth advertising to you].