Wireless Questions

ShortyShorty Manchester, UK Icrontian
edited August 2003 in Science & Tech
I need information and assistance from the gurus (prime?, Thrax?) :(

Where I work, we are in a situation where we are going to be moving partition walls to increase lab space. Not a problem, except about 30 RJ-45 ports serving voice and data. We got contractor quotes in to move/extend them but it touched nearly £4000 :eek:

So... my CEO has decided to go wireless everything. That's 30 LAN connections and 30 DECT phones :eek:

802.11G I believe is the fastest (54mps). Anyone have experience? Im thinking of getting 3 wireless access points (front office, backoffice and lab). This will cover the whole building effectively. I have very little knowledge of wireless & security.

Tips? Hints? Professional opinions? Guidance? :bawling:

Comments

  • ThraxThrax 🐌 Austin, TX Icrontian
    edited August 2003
    You have a few ways to go about this:

    802.11A, although a dying standard, operates in the 5GHz band. Because of this higher frequency it has an easier time punching through walls and other solid surfaces for good transmission rates. It's 54Mbps, but only other A devices can work with it. Not many things operate in this frequency. The equipment is expensive and unique.

    802.11G, is the blossoming standard. Ratified at LOWER speeds than 54Mbps to make sure customers really do get what they buy (The lower numbers are more accurate. I believe it's in the 30MB/s range). 802.11G is compatible with both B AND G devices, as it operates in the 2.4GHz band. In that same band are many cell phones, cordless phones, bluetooth, and a whole damn lot of other wireless technologies...But the interference is minimal in most cases. The 2.4GHz band also doesn't have as much punching power to go through solid surfaces. With this, assuming the clients don't need a maximum bandwidth of 5.4 megs/sec, you can outfit the client stations with 802.11B equipment to save money. Outfit the high-bandwidth stations with G equipment.

    Also, another avenue to pursue is the new tri-band gear pioneered by Linksys. It incorporates 802.11A/G chips and can be toggled between 5GHz and 2.4GHz. This might be beneficial as devices behind several walls or hard surfaces can use 802.11A gear and transmit to a unique subnet running the A/G gear in A mode. The rest of the subnets broadcasting in the clear could use B cards transmitting to 802.11B-mode WAPs.

    Security!

    Assuming you have three networks as you said:


    Unique 13-25 character ASCII WEP code (Wireless equivalent protocol).
    MAC filtering. Edit the access lists in the WAPs to allow connections only from MAC addresses of computers at this location.
    Give the WAP a unique frequency. 2.401-2.499 can be selected on many models.
    Give the WAP a unique password.
    Give the WAP a unique SID (Service ID).
    Give the clients the unique WEP to broadcast.
    Make the clients broadcast on the unique frequency.
    Make the clients connect to the unique SID to their network.


    That'll make it so only clients transmitting the right WEP, with the right MAC, the right frequency, and broadcasting to the right WAP can connect.

    Do the same for each group of computers. Make all of them broadcast to authorize with different WEPs, frequencies, and SIDs.


    Example:

    Front office:

    <b>WAP Settings</b>
    Frequency: 2.402GHz listen
    SID: FrontOfficeWAP1
    WEP: FrontOffice2358
    WAP Pass: Whatever you please
    MAC filter: Only MAC addresses from clients in the front office

    <B>Client Settings</b>
    Frequency: 2.402GHz broadcast
    SID: FrontOfficeWAP1
    WEP: FrontOffice2358
    MAC address: One of the listed MAcs in the WAP


    <b>Final Words:</b>

    802.11A/G have a maximum of 54Mbps. Typical throughput is expected in the 22-28Mbps range, and about 4-6Mbps for 802.11B. Netgear's turbo mode ups the throughput to about 26-32Mbps, and Proxim's turbo mode ups it again to 35-42mbps...Not really as advertised.

    Higher bandwidth = lower coverage radius.
    Lower bandwidth = Higher coverage radius.

    The faster the speeds, the higher the frequency, and the quicker the device cuts the signal off. Wireless equipment have frequency times. Once a certain number of full wavelengths are sent out or received, it starts another packet. So if you have a higher frequency, the waves won't go as far to and from the device, but the bandwidth is higher.

    You strike the balance.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited August 2003
    All I can offer is that Proxim's 802.11a equipment includes a proprietary turbo standard (not the same as netgear's) that bumps the throughput from 54mb/s to, no... not 74mb/s (which is what the netgear does) but to 108mb/s

    For a wireless connection, 108mb/s is FAST
  • primesuspectprimesuspect Beepin n' Boopin Detroit, MI Icrontian
    edited August 2003
    In my experience, wireless performance is never what you quite expect. While bandwidth usually tops out at the spec (54mb for a/g, 11 for b), actual throughput is noticeably lower than wired equivalent. Also, if security is truly a concern, then please understand that wireless security is pretty much a joke. Anybody who wanted to get in, could. Thrax's idea of using MAC filtering is a good one, but as you know in a production environment, maintaining a list of authorized addresses can quickly become a full time job in itself. I have found the wireless is about 98% percent ready for enterprise, but it's just not quite there yet. You are asking for a world of aggrivation if you have to go wireless, then you have to, but I would actually recommend against it for actual production usage. Your job will become a whole new level of frustrating.

    Wireless in the enterprise is excellent for providing mobile users a quick way to connect to the corporate LAN, and to provide on-site customers with internet access, but for desktops, it's just not reliable enough. Trust me dude, while theory holds well in this case, I can say from actual rollouts and working experience that 30 desktop computers on a wireless lan is going to be a gigantic nightmare. The TCO on wireless is sky-high, almost entirely due to support and maintenance. Whereas the TCO of extending the cable is going to be lower, because once the job is done, maintenance issues will be nil. Wireless networks require active maintenance, so you will be adding a new bullet point to your job description.

    Is it set in stone? The wireless thing? If it is, best of luck to you, and of course we'll all be here to help you out as much as possible, but my experience tells me to highly suggest to your boss to rethink the wireless concept for desktop computers.
  • ShortyShorty Manchester, UK Icrontian
    edited August 2003
    Thank you Guys :)

    This is just the kind of information I desperately need.

    My hands are totally tied :( I don't have the option to choose.

    prime remembers my recent phonecalls regarding the lack of understanding with my superiors on networking issues. Im gonna try to persuade them, that this is a bad idea. Im just hoping they see sense. Thank you for the enterprise analysis. It's gonna be my argument agains't the idea :)

    Thrax, Thanks mate for the tips and assistance. I figured you would be a good person to speak to, as you have dealt with wireless and have a strong understanding of the core underlying technology :)

    Geeky, I was looking at that kit too. However, my budget is so damn tight, it's gonna be hard to get them to take the "turbo" option :( and it's price "premium".

    This is a logistical nightmare. The more I have read and understood the information presented, the more I am starting to have serious concerns over speed, security and deployment.

    ~sigh~ Some really long days ahead for me :(
  • reelbigfishreelbigfish Boston, MA Member
    edited August 2003
    In my experience with wireless, 802.11b, I have problems playing games on a desktop. Therefore, I doubt wireless would be reliable enough for voice traffic. It's fine for web browsing, but if you constantly need to transfer data, the lag may be too much.
  • ThraxThrax 🐌 Austin, TX Icrontian
    edited August 2003
    I game regularly on a single 802.11G router handling 3 computers with gaming traffic. The closest client machine is 5 feet away, the furthest one is through a solid oak floor, then solid oak wall approximately 16 feet straight up.

    I don't have lag problems. But then again, it's not B traffic. :)
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited August 2003
    Um, wait for A. Unless you want either a repeater or an access point on any room of stone or block or brick walls or very thick insulated wood or every room built with steel wall risers. The current fastest that is more economical does not have a good wall panetration freq set avlaibale to it. This makes it both the most secure and frustratingly the least able to work well in a many-roomed "home." One way around this is a mixed-mode net, with multimedial (wireless plus hardwired access points that are hard wired to each other or to a trditional switch to eliminate the extreme signal to distance and density loss of the current fastest economical wireless). Probably it is also the best economical way to deploy at first too, to see where you DO get signal loss in reality. Switch a room over, then use a laptop and measure rates from other rooms. A fiber backbone would be best right now, next best would CAT 6 Sheilded, or CAT 6 Plenum -- your backbone will determine best flow from one end of faciltiy to the other, and G's is weak due to signal attentuation as opposed to base speed. A's has a security problem, it might be a tib too robust unless your site has very thick and dense outer walls.

    So, would tell your boss that things need to be staged for this kind of change to balance the need to keep private code private and have good speed that is relaible. I have known folks to tell me that they had to buy 4-5 times as many repeaters with G deploys as they figured on. These folks work in hospitals in my area, and they need secure installs with very relaible data flow-- they run gateways wired to remote antennae in each room and every 50 to 100 feet in halls for laptop or remote terminal use in rooms and in halls, and the backbone is CAT 6 Plenum for long runs with critical flows and CAT5E for non-critical run with Fiber for anything from building to building.

    I am not a wireless guru per se, but have talked to very heavy site admins for 3-4 story hurricane hardened hospitals and toured hospitals with them. They have been known to wire admin with G and other places with A just to handle interference and other problems, but most of them use G with lots of romm and zone antennaes hooked to repeaters and gates that concentrate repeater signals, up in drop ceilings. Where they did not have dropped ceilings they used CAT 6 or fiber depending on priority and distance and hickness of walls. One gate should prove this, with a laptop to show hoe signals actually decrease in your facility. "sniff it yourself," for G, is the best way to go. You could do two laptops, put one gate on a cart, move from room toroom,and see where signal dropouts are, with BSD or Linux if the laptops are speced right, and with one desktop on a cart and the gate on same cart if the budget only runs to one laptop.

    TECHNICALLY, you could use TWO carts and two desktops(one per cart), but a big UPS that could power a small basic PC fora couple hours might be a problem to get on the moving and sniffing set of gear. If you have to, tell him a real good security survey would cost more than one laptop. It would be quite true.

    A good G-capable PDA that can hotswap main batteries with 5-6 spare batteries (rechargable) would substitute for a laptop for signal propagation mapping along with plans of building, laptop would be more robust to network troubleshoot the whole thing in the long run.

    I do not know if a Fluke Netrunner can be made G-capable without a gateway for itself and a UPS to power the gateway. I DO know those are "worth their weight in Gold" for network fixing very quickly and with less network-trained folks. A Netrunner is a network mapping thing, that can measure and show signal loss. If you can beg borrow, rent, use one for a day with a G gateway and an A gateway and a lappie with a couple different media wireless things and drivers, you can range-and-signal-loss-map the building in about 12 hours of work unless it is radically huge and that will save him very possibly doubling his costs if no such thing is done and you do it as you deploy, due to rework and extra gear to fill in "holes."

    I am on the wrong continent to do that job,but you NEED a signal propagation map and having one will let you tell exactly what is up if you have a laptop or wireless capable device to verify map with later and extend or change it as needed.

    Good Luck.
  • maxanonmaxanon Montreal
    edited August 2003
    What is the price difference between the wireless and the re-wiring?

    How difficult is it to wire it yourself (with some friends)? What is the cost of just the materials? I've done small businesses and homes and it wasn't too difficult. Definitely less of a headache than wireless.

    It sounds like your boss likes to read PC mag (or some equivalent) and they think that they're immediate IT gurus. The place that I work has evaluated it (wireless) and deemed it not secure enough. Physical security is the best security.

    You can show him articles http://www.wardriving.com that will cause a pause for concern.
  • ShortyShorty Manchester, UK Icrontian
    edited August 2003
    Thanks for the continued replies :)

    Maxanon, There is no real price difference. The re-wiring professionally would cost £3500. The total wireless solution would cost £3000.

    The problem is just as has been analysed. My CEO considers himself to "IT literate". I have no option. If he says "go wireless", we go wireless. I've made my reservations well known, I've dug out the DECT whitepaper for phones and studies on contended wireless LAN.

    Did it make a difference? Not in the slighest. I've suggested we rewired in-house. That didn't get taken up either. He seems insistent (which means he thinks it's a great idea, so the rest of us have to agree).

    {Sigh} .. the wireless kit and phones are ordered. I start building it tomorrow. Wish me luck, I am going to need it :(
  • NecropolisNecropolis Hawarden, Wales Icrontian
    edited August 2003
    Dan,

    For the right price I could call some people and they will remove you problem for you. You wont miss you boss will you :D

    I know the feeling tho, they always know best.
  • maxanonmaxanon Montreal
    edited August 2003
    Good luck. I hope it all goes really smoothly.
  • ShortyShorty Manchester, UK Icrontian
    edited August 2003
    ;D;D;D;D;D;D Today RULED.

    I got us a quote from the guys doing the partition wall moving.

    £900. Yyyyyeahhhhhhh!!!

    So.. I place the delicate information on the desk and leave quickly. My CEO is none too impressed. This effectively kills the wireless idea stone dead :D

    He is still "considering" it, but this brings hard wired LAN back into the picture. Quick smart :cool:
  • NecropolisNecropolis Hawarden, Wales Icrontian
    edited August 2003
    Great Stuff :thumbsup:

    YAY for dodgy British Builders ;D;D;D;D;D
  • ShortyShorty Manchester, UK Icrontian
    edited August 2003
    Necropolis_uk said
    Great Stuff :thumbsup:

    YAY for dodgy British Builders ;D;D;D;D;D

    Im gonna take some photos of the work as it is being done :D

    Heehehe.. they are dodgy mate but they will save me the pain of wireless!
  • NecropolisNecropolis Hawarden, Wales Icrontian
    edited August 2003
    Shorty said
    Im gonna take some photos of the work as it is being done :D

    That should make for some interesting viewing. 40 Pictures of men having tea breaks ;D
  • ThraxThrax 🐌 Austin, TX Icrontian
    edited August 2003
    ;D
Sign In or Register to comment.