[NA] LAN cabling Question

[AggieMike88]

I'm renovating the office and front public area as part of my call center upgrade.

Part of this is having the electrician come out to change/improve the electrical service. The electrician can also handle pulling data cable and installing the proper terminations so that each station has a connection to the LAN and there is a central "hub" point for the server, switch, router, and all.

For each workstation, I need to connect three devices to the LAN: PC, VoIP Phone, and Printer.

The original design I'm working under was before the small gigabit switches were plentiful and affordable. So there are a wall boxes all over the place that have multiple plugs, each with one wire running back to the central gang terminal. A big mess and a real nightmare to manage.

To simplify, can I just run a single wire from the central terminal to each station, and then use a gigabit switch as a splitter? Or is the old style "one cable per device" the way to go?

 

[EppyGA]

You can use the switch

 

[Checkout_my_Six]

Wifi is your friend....unless the data needs protection.

 

[James331]

Wifi is your friend....unless the data needs protection.

Unless you have metal in the walls, brick, need better speed, security, etc.

If you're pulling wires anyways, it's stupid to not run Ethernet cable, it's a tech that ain't going anywhere, another thing I'd do is install some of the power outlets with USB chargers on then, for places next to chairs and couches.


To OP, yeah run it all to a switch or hub depending on what you're doing.

 

[ifly4fun]

Not sure how big your lan will grow - but there is benefit to keeping them separate. If you decide to keep the Workstation/VoIP/Printer all on seperate VLANs, you would then need a managed switch at each workstation to do the VLANs. This would probably negate any savings of not running the cables. VLANs can help you with security and better traffic management (QoS, etc..). Typically the expense is the labor of pulling the cable, and the labor should be the same regardless of 1 or 3 cables (or negligible). CAT6 is cheap these days too.

 

[weilke]

I like VOIP phon a separate wire as I have them on a separate router. Other than that, gigabit has so much volume that you can run 1 wire and stick a 5-port switch under each desk. Just secure the powe connection so your CSR doesn't unplug that superfluous power wart so she can plug in her ipad (or use a PoE switch in the closet and use PoE powered satellite switches).

 

[sferguson524]

Most VOIP phones have a small 2 port switch built into them You define a voice VLAN at the switch, and the phone's switch does the magic.

 

[jesse]

I'm renovating the office and front public area as part of my call center upgrade.

Part of this is having the electrician come out to change/improve the electrical service. The electrician can also handle pulling data cable and installing the proper terminations so that each station has a connection to the LAN and there is a central "hub" point for the server, switch, router, and all.

For each workstation, I need to connect three devices to the LAN: PC, VoIP Phone, and Printer.

The original design I'm working under was before the small gigabit switches were plentiful and affordable. So there are a wall boxes all over the place that have multiple plugs, each with one wire running back to the central gang terminal. A big mess and a real nightmare to manage.

To simplify, can I just run a single wire from the central terminal to each station, and then use a gigabit switch as a splitter? Or is the old style "one cable per device" the way to go?

You generally try to avoid this as much as possible. The most desirable thing is that every ethernet port, in your whole building (if physical dimensions permit), terminate to individual ports at one central location. This allows you to have a nice managed switch that can do all sorts of nice network things for you.

If you start dropping switches all over the place you lose the ability to have one nice central place to manage your network (managed switch) and you have a lot of **** that can fail all over your network

I would recommend that at minimum each workstation has a network port. All of those network ports should terminate at one central location on an ethernet patch panel. Cables then run from that patch panel to your switch. If you go this route you'll need VOIP phones that have a switch built in. Most of them do and they usually support Vlans. Probably not something you'd mess with at the size of your office but it is a nice feature. (Allows you to have your VOIP traffic on a seperate "virtual" network then your LAN traffic. makes it easy to use QoS).

That said I consider the above the minimum. Typically if budget permits I want two network ports terminated at each workstation all running back to the central location. This just future proofs things better, IMO.

Don't forget about ports for your printers. Once again do try to avoid using switches all over your network. The more you can stick to one nice big managed switch the better.

At my last gig we had a good sized building and a good 500 ethernet cables terminating in one central network closet. Every workstation had 2 ports. Some offices had as many as 4. That was a very expensive cabling job. We never added cheap switches anywhere. We'd pull new cable all the way from that central location before we'd add a switch anywhere.

 

[jsstevens]

You generally try to avoid this as much as possible. The most desirable thing is that every ethernet port, in your whole building (if physical dimensions permit), terminate to individual ports at one central location. This allows you to have a nice managed switch that can do all sorts of nice network things for you.

If you start dropping switches all over the place you lose the ability to have one nice central place to manage your network (managed switch) and you have a lot of **** that can fail all over your network

I would recommend that at minimum each workstation has a network port. All of those network ports should terminate at one central location on an ethernet patch panel. Cables then run from that patch panel to your switch. If you go this route you'll need VOIP phones that have a switch built in. Most of them do and they usually support Vlans. Probably not something you'd mess with at the size of your office but it is a nice feature. (Allows you to have your VOIP traffic on a seperate "virtual" network then your LAN traffic. makes it easy to use QoS).

That said I consider the above the minimum. Typically if budget permits I want two network ports terminated at each workstation all running back to the central location. This just future proofs things better, IMO.

Don't forget about ports for your printers. Once again do try to avoid using switches all over your network. The more you can stick to one nice big managed switch the better.

At my last gig we had a good sized building and a good 500 ethernet cables terminating in one central network closet. Every workstation had 2 ports. Some offices had as many as 4. That was a very expensive cabling job. We never added cheap switches anywhere. We'd pull new cable all the way from that central location before we'd add a switch anywhere.

I've supervised three space buildout in the last three years. Every network drop has 3 or 4 CAT 6 cables to the closest network closet. I've never had anybody complain that "There's too much network here." Installing extra cable or two is cheap when you're already installing 1. And multiple switches without a good configuration can cause all sorts of nasty, intermittent network problems. Things will work one day and not the next. Power it all down and back up and voila. Until next time. Just don't go there.

John

 

[flyingron]

We always ran three to each desk: One for the main data network, one for the phones (believe me, you've never heard grumbling in your life until some data farkup swamps your business VOIP service), and one for an "alternate" data network.

 

[Ravioli]

Home run those cables. Three per per work spot. (hopefully your space has a good spot to put the data center. Label them well. (Sounds like a dumb thing to skip, but if you've ever been there, you'd know)

 

[cowman]

Cables are cheap so unless there's some reason not to run so many cables, might as well run as many as possible. Heck put a couple extra ports in for expansion in the future.... and by god label them at both ends.

That said my home setup is a switch and a bunch of devices in the house, then a single gigabit line up to the workshop(well actually 2 in case one ever fails) and another switch up in the workshop. Unless you have a lot of employees working simultaneously or some super high bandwidth application going this should be absolutely fine for typical small office document sharing/printing type activities.

 

[Ravioli]

Before you run them cables... why not all wireless? You can do wireless VOIP phones, computers, etc... save a bunch of wiring, add a bunch of flexibility.

 

[sferguson524]

Before you run them cables... why not all wireless? You can do wireless VOIP phones, computers, etc... save a bunch of wiring, add a bunch of flexibility.

One issue I have with wifi is it's largely first come first served for data at the AP until you get into MU-MIMO. Not to mention RF interference from neighboring businesses etc. The 2.4 gHz spectrum is getting quite crowded, and 5 gHz doesn't have the penetration

 

[Ravioli]

One issue I have with wifi is it's largely first come first served for data at the AP until you get into MU-MIMO. Not to mention RF interference from neighboring businesses etc. The 2.4 gHz spectrum is getting quite crowded, and 5 gHz doesn't have the penetration

This is for an auto salvage company. My mind's eye sees that as a large facility that is unlikely to have considerable neighboring traffic. I probably agree with you if this was a more traditional setting.

 

[Let'sgoflying!]

I just did this at my (small) business Mike and decided against wifi (never considered workstation switches but would not have done that either). Home run cables just seemed more stable. Rip out all that old wiring, use it to pull new cables. Much easier to work in the attic now at 80F than during the peak of the summer!

 

[cowman]

Wifi either works well or it doesn't. Depends on the devices, the structures, and the environment but there's little you can do about that if you end up with a finicky connection.

I'd always prefer wired if practival for that reason.

 

[John221us]

You probably have lots of extra wire you can strip from those cars

 

[AggieMike88]

While we do have wireless service as a feature of our router, it isn't something I want to depend on for second-by-second operations.

 

[Ravioli]

This part is kind of like plumbing. At some point you're relying on the connection from your Inbound Demarcation to their network. Wire or wireless, your fail point is the same.

 

[denverpilot]

Others covered the desktop switches. Don't do that.

As additional reason why: If you decide to install power backup for critical things, you probably don't want to be powering a bunch of desktop switches.

But then again, if you're using desktop PCs you'll have to decide if cheap desk UPSes are in order for some functions.

You definitely want to think about Power over Ethernet for the phones. Same reason. Power them from the switchroom and someday when you decide to battery or generator power the swithroom, the phones come along for the ride. With appropriate amounts of backup power for them.

 

[jsstevens]

Others covered the desktop switches. Don't do that.

As additional reason why: If you decide to install power backup for critical things, you probably don't want to be powering a bunch of desktop switches.

But then again, if you're using desktop PCs you'll have to decide if cheap desk UPSes are in order for some functions.

You definitely want to think about Power over Ethernet for the phones. Same reason. Power them from the switchroom and someday when you decide to battery or generator power the swithroom, the phones come along for the ride. With appropriate amounts of backup power for them.

POE VOIP phones are very nice. They are easy to move. Thy are easy to provide emergency power for (as Nate said). Properly configured they are really easy to manage.

We now support 4 geographically diverse offices on the same phone system. It's all network packets. (With proper QOS and VLANs of course.)

John

 

[deonb]

Others covered the desktop switches. Don't do that.

Funny. I worked in a 50'000 office environment where each and every individual office had an indidividual desktop switch.

It was probably the only part of their IT infrastructure that I liked. When the main network was down I could still access my printer and other computers in my office.

Of course, I wasn't the one who had to maintain that stuff.

 

[denverpilot]

Funny. I worked in a 50'000 office environment where each and every individual office had an indidividual desktop switch.

It was probably the only part of their IT infrastructure that I liked. When the main network was down I could still access my printer and other computers in my office.

Of course, I wasn't the one who had to maintain that stuff.

The fix for that is not to let the "main network" go down. Ha. What the hell would they be allowing that for, more than maybe once every four or five years to upgrade devices at night when the place is quiet or closed, or maybe a core router update, which with appropriate money, can also be done with no significant downtime?

50,000 users and they let their network go down regularly. Someone needs to be slapped.

 

[denverpilot]

POE VOIP phones are very nice. They are easy to move. Thy are easy to provide emergency power for (as Nate said). Properly configured they are really easy to manage.

We now support 4 geographically diverse offices on the same phone system. It's all network packets. (With proper QOS and VLANs of course.)

John

Yup. We love our PoE. The VOIP phones have their own VLAN and auto-configure over it. Folks still have that weird "Hey can you schedule to move so-and-so's desk?" habit of being ultra polite from the old days and old phones. We just laugh and say, "We can walk over and pick up her phone and move it if you like, or she can. Just plug it into the blue jack labeled Voice and a number under the desk. Plug the PC into the orange one labeled Data and a number."

(We cheaped out and a FEW desks have no PoE on the Data jack so it's just easier to explain to plug into voice but for 90% of the desks it doesn't matter.)

When we re-did the wifi access points the PoE made our lives a whole lot easier too. Just run a cable to the place on the ceiling the AP needed to go and the APs all power off of the switchroom also. No need for power drops on the ceiling. Since it's a high ceiling warehouse building, the APs are bolted to the cable trays and hanging underneath them overhead.

 

[Ravioli]

Perhaps 50,000 sqft?

 

[John221us]

Funny. I worked in a 50'000 office environment where each and every individual office had an indidividual desktop switch.

It was probably the only part of their IT infrastructure that I liked. When the main network was down I could still access my printer and other computers in my office.

Of course, I wasn't the one who had to maintain that stuff.

Rogue desktop switches are terrible idea. It hinders troubleshooting, decrements security, doesn't allow port management and makes delivering PoE very difficult. They happen in most big environments, but it is usually someone being lazy and in a hurry to add a printer or something like that. Some environments I have worked, adding a rogue switch could get you fired. Rogue Access Points are another big issue...

 

[deonb]

Perhaps 50,000 sqft?

50'000 offices.

In the group I worked on the main network fairly regularly went down since we worked on creating networking software & hardware and we were our own beta tester.

But the desktop switch per office policy applied to the entire company. Most people had an 8 port switch, with I guess an average of 6 devices in each office that required networking. I had a 32 port switch in mine filled with various stuff (including more switches ).

At some point each office had 2 active taps - one for network, one for analog phone (with a CAT5 port - which is fun if you accidentally plug an Ethernet device into it). But after the switch to VOIP we just went to the single CAT5 + desktop switch. I couldn't be bothered with the VOIP phone - it had an absolutely beyond horrible UI to enter your username & password, which would take me about 5 minutes, and it would log you out every second day. So I just shoved in on a shelve and told people to call my personal cell instead.

But yes, in our company we didn't have a lot of respect for the IT department. Especially since most people knew more than IT, but IT wasn't particularly trained to deal with us.

So MANY a phone conversation with them went like this:

"Your server is down. Can you please reboot it."
"How do you know it's not your machine?"

"Because I can't reach it from any of the 5 machines in my office."
"Still... try rebooting your computer."

"Which one?"
"The one that you're trying to reach it from."

"It's from 5 machine... oh never mind. Let me try.".
Get up, walk to the break room (remember I use my personal cellphone). 2 minutes later - after not being anywhere near my computers now: "I rebooted it - same problem".
"Try pinging it by the IP address".

(While making coffee - phone now squeezed on my shoulder). "Ok, just tried that. Still no luck"
"Have you tried from your neighbor's office. Maybe it's your part of the network that's down".

"I can connect to the server right next to that one."
"Well, try from your neighbor".

"I already tried my neighbor before I called you. I then contacted someone in a different building. Then I connected over my cell and VPN'd back into a completely different network. It's your @#!$$ server that's down!"
"I don't see an outage report."

"This is me filing the outage report. Can YOU ping that server?"
"I can't connect to that server in general - it's private"

"You should still be able to ping it. It's not on IPSEC."
"Mmm... I can't ping it... Ok, I'll report it as out."

"Nevermind, the server is back up now."
"Glad we could help! Let us know if you have ..."

<click>

 

[denverpilot]

At some point each office had 2 active taps - one for network, one for analog phone (with a CAT5 port - which is fun if you accidentally plug an Ethernet device into it).

Not sure why. Any modern Ethernet device worth purchasing can isolate both the 48V DC on-hook voltage and the 90V RMS AC "ring" voltage on pair one (pins 4 and 5) and probably on pair two (pins 3 and 6) just fine. Anything on the market that can't, isn't worth buying. What garbage were you using? Nothing modern cares.

As far as the rest of it goes, sounds like the organization didn't have any respect nor understanding of what an infrastructure group does, or offers, and just kinda slapped some people at a "help desk" and pretended to have a clue. But that's pretty common in large organizations these days. Especially organizations full of engineers who the bosses think build similar stuff to what an infrastructure group would USE but doesn't bother to figure out which sorts of people and methods to apply to the infrastructure itself. Once there's no quality control of infrastructure and metrics for uptime because someone thinks their user group can manage it better than those tasked to do that job and someone lets them fart around unprofessionally, yup... it'll always be broken.

Example: No place that takes their infrastructure seriously does beta tests with the main infrastructure. That's just stupid. There's a reason your VOIP phone "rebooted every other day" - it's because the retards didn't treat infrastructure as separate from your test network. You don't do lab work on the business' main network.

It's beyond stupid to spend capital on a VOIP phone system that doesn't work because unqualified people are constantly dicking with and breaking the network that's supposed to support them.

 

[AggieMike88]

Others covered the desktop switches. Don't do that.

As additional reason why: If you decide to install power backup for critical things, you probably don't want to be powering a bunch of desktop switches.

But then again, if you're using desktop PCs you'll have to decide if cheap desk UPSes are in order for some functions.

You definitely want to think about Power over Ethernet for the phones. Same reason. Power them from the switchroom and someday when you decide to battery or generator power the swithroom, the phones come along for the ride. With appropriate amounts of backup power for them.

Not sure if the Polycom VVX series does PoE... will need to investigate since that would reduce the rat tails clutter by one wire.

Maybe I should just send you a SWA ticket and tell you to pack your crimpers, and then hire you as the contractor for this task.

 

[denverpilot]

Not sure if the Polycom VVX series does PoE... will need to investigate since that would reduce the rat tails clutter by one wire.

Maybe I should just send you a SWA ticket and tell you to pack your crimpers, and then hire you as the contractor for this task.

Sure they do. In fact... there's more to it than that if you have "sidecars" or other high power devices plugged into their USB ports.

http://support.polycom.com/global/d...ts/voice/Power_Consumption_and_Management.pdf

(You realize even though I didn't deal with their phones as much as larger gear, that's who I worked for forever, right? Good old PLCM... stuck at half their market value since around the time I left. They went up enough that my old boss and good friend's options weren't underwater in almost a decade and he finally got to cash out. I had one of the very first VVX models on my desk. Giant thing with a huge video display. The new stuff is totally different now.)

As far as traveling for that biz goes, it'd take a lot to get me to do that again. But I'd rather just fly the 182 down. And it's in the shop right now.

It's pretty rare to not have PoE on VOIP phones these days. It's too useful.

 

[John221us]

Not sure why. Any modern Ethernet device worth purchasing can isolate both the 48V DC on-hook voltage and the 90V RMS AC "ring" voltage on pair one (pins 4 and 5) and probably on pair two (pins 3 and 6) just fine. Anything on the market that can't, isn't worth buying. What garbage were you using? Nothing modern cares.

As far as the rest of it goes, sounds like the organization didn't have any respect nor understanding of what an infrastructure group does, or offers, and just kinda slapped some people at a "help desk" and pretended to have a clue. But that's pretty common in large organizations these days. Especially organizations full of engineers who the bosses think build similar stuff to what an infrastructure group would USE but doesn't bother to figure out which sorts of people and methods to apply to the infrastructure itself. Once there's no quality control of infrastructure and metrics for uptime because someone thinks their user group can manage it better than those tasked to do that job and someone lets them fart around unprofessionally, yup... it'll always be broken.

Example: No place that takes their infrastructure seriously does beta tests with the main infrastructure. That's just stupid. There's a reason your VOIP phone "rebooted every other day" - it's because the retards didn't treat infrastructure as separate from your test network. You don't do lab work on the business' main network.

It's beyond stupid to spend capital on a VOIP phone system that doesn't work because unqualified people are constantly dicking with and breaking the network that's supposed to support them.

This is 100% correct. I have been in large engineering environments where the engineers "knew more than IT". They may know more than the help desk guy asking him to reboot, but they don't know more about supporting large networks. Basically, when you see this stuff (rogue switches, rogue servers, etc.) IT has given up. Without standardization, lockdown policies and end to end management, the level of support and network stability is going to be poor. Go ahead and blame IT, but the reality is that you are testing in production and creating your own world. Don't expect a level one help desk person to be able to unwind that mess. If you can't contact the server it could be any number of issues, but with three unmanaged switches stuck between you and the server, good luck figuring it out.

 

[John221us]

Not sure if the Polycom VVX series does PoE... will need to investigate since that would reduce the rat tails clutter by one wire.

Maybe I should just send you a SWA ticket and tell you to pack your crimpers, and then hire you as the contractor for this task.

The VVX series is a great phone. They definitely support PoE. The can be centrally configured and managed and offer a lot of great features, like intercom and paging.

 

[deonb]

Not sure why. Any modern Ethernet device worth purchasing can isolate both the 48V DC on-hook voltage and the 90V RMS AC "ring" voltage on pair one (pins 4 and 5) and probably on pair two (pins 3 and 6) just fine. Anything on the market that can't, isn't worth buying. What garbage were you using? Nothing modern cares.

This was in the 90's. Days of 3c509's and NE2000 clones.

As far as the rest of it goes, sounds like the organization didn't have any respect nor understanding of what an infrastructure group does, or offers, and just kinda slapped some people at a "help desk" and pretended to have a clue...

The IT department was a completely separate division under a different VP, having only IT as a charter. In fact they got often cited in case studies on how to do things, and eventually they started bringing in revenue by themselves (outsourcing their services to other companies). So they must have done something right I assume... but they were hated by everybody I knew.

Not because the services didn't work - they worked for the most part. Outside the limited test network (which only affected a small group of us), it was more stable than anywhere I've seen since. But when you phoned them up you inevitably get someone that know way less about things than you do that you have to go 10 rounds with before you can speak to someone who can do something about your problem. That was particularly painful - we felt that model wasn't a fit for a tech company.

Example: No place that takes their infrastructure seriously does beta tests with the main infrastructure. That's just stupid. There's a reason your VOIP phone "rebooted every other day" - it's because the retards didn't treat infrastructure as separate from your test network.

It wasn't the infrastructure. The way the VOIP timeout thing was explained to me is that the phone had a 48 hour timeout if you didn't use the phone. And in fact you were really supposed to sign out every day when you left. So that "Other people can't come into your office when you're not there and make international phonecalls". Except we had absolutely no requirement against that. Nobody could care less if employees / visitors / janitors make 2 hour international phonecalls. It was a requirement that the IT department dreamt up by themselves.

They eventually switched to different phones that log in automatically when you log into the PC. Which I also didn't bother with since virtually the only reason for me to ever use the phone was to call the IT department when I can't log into my PC... I think in 17 years I used my desk phone maybe a total of 30 times, and 25 of those times it was to call IT. A phone to me is the thing you use when you want to report problems with your real communication system.

 

[denverpilot]

This was in the 90's. Days of 3c509's and NE2000 clones.

It wasn't the infrastructure. The way the VOIP timeout thing was explained to me is that the phone had a 48 hour timeout if you didn't use the phone. And in fact you were really supposed to sign out every day when you left. So that "Other people can't come into your office when you're not there and make international phonecalls". Except we had absolutely no requirement against that. Nobody could care less if employees / visitors / janitors make 2 hour international phonecalls. It was a requirement that the IT department dreamt up by themselves.

Hmm... 50,000 user company doing VoIP in the 90's with phones that could do user auth. Guess that narrows it down.
And an IT department that outsourced to other places. Narrows it even further.

You wouldn't catch me dead working for any of the three names that brings to mind. They are sweatshops for IT infrastructure people. Literally. Two were good places to get a start and then GTFO, the third was so bad nobody with a brain went there.

And the rest of the story keeps leaning toward what I would categorize as what probably really was going on, "Spoiled engineering group that was annoying enough that they're getting BOFH'd, permanently." I've seen that a number of times inside companies of that size. Kinda funny that they never catch on, too.

The logout thing was probably not "dreamt up by IT", they probably had to meet a stupid (even retarded, yeah, I really mean that word) security certification standard that required the phones be kicked out every 48 hours. Just so some doofus could claim they met the standard to pitch that as part of how wonderful they were to buy outsourcing services from.

(It's rare that a security certification that actually contributes significantly to real security, but boy do they contribute to annoying the holy hell out of internal customers. But once someone gets a bug up their butt or reads about some security cert in "CEO Magazine" or where ever they heard about it... men's room at the golf club, who knows? -- insert punch to someone's face at Gartner Group, especially back in the 90s, here too -- and decides they want it, no expense will be spared and no user experience will be cared about ever again, if it breaks the rules of the security cert.)

 

[Escher]

Well, I guess this is as good a use for a first post as any.

Background: Worked a career as a network engineer. Cisco certified at the CCNP level, Certified Wireless engineer, etc. I've done build-outs, re-furbs, govt & enterprise support, project management for Fortune 20 companies and even spent a couple of years doing carrier level VoIP engineering where I worked on BGP speakers for the Tier 1 provider with the "three in a row" 700 series AS numbers. (That's an inside baseball secret handshake in case there's another network nerd around.) In plain ENGLISH, I could turn the internet on and off.

To the question at hand, here are some thoughts...

1. Wireless is great. If you have an actual wireless engineer who knows what he's doing and are willing to spend the money to do it right. If your workspace involves a lot of mobile computing, whether it's laptops or hand-held scanning devices in a warehouse, do it. And use real enterprise level equipment. Otherwise limit your wireless use to guest access, preferably on a completely separate link. As in for instance, a Verizon MPLS-VPN for your production network and a Time Warner cable drop for your wireless guest network.

2. It's one thing to have a "star" style wiring layout with a main network closet housing a Cluster Commander switch and sub-closets with slave switches. You can limit access and they are actually designed to work that way, including things about the design that make troubleshooting easy. Running wires to switches that can be reached by end users instead of home running every network port does several bad things. When troubleshooting it takes a single point of failure and turns it into at LEAST five, assuming the smallest desktop switches sold. You also lose control of what gets plugged in. Never mind silly stuff like people using the wrong ports, I'm talking about people deciding to create their own personal wireless LANS. A while back I worked for a good sized school district where about 70% of our service call time was spent on un-frakking something someone had done to the publicly accessible switch. And not a day went by when the network in one of our 130-ish school buildings went didn't go down and we subsequently found devices with 192.x.x.x IP addresses on our 10.x.x.x network because some teacher thought he/she was above the rules and went to CompUSA to buy a wireless router and then plugged it into our network without knowing what they were doing. Suddenly there are two devices on the same network handing out mutually incompatible IP addresses. And computers, being marginally more intelligent than a rock, take the first IP address that comes back in response to it's request for one.

3. In addition to the aforementioned difficulties in troubleshooting and security, the routers and switches that you can buy as a "civilian" as opposed to a professional vendor authorized to sell the good stuff, are not actually routers and switches. This kind of gets into the weeds so don't worry about it, but the switches you can buy at the big box store are just Hubs that can separate collision domains and the routers are switches than can do basic LAN routing, usually using RIP1, RIP2 and OSPF. The difference between consumer equipment and professional equipment is in a certain sense subtle but it comes down to the sophistication of the decisions the equipment can make, how fast it can make them and how many little ONES and ZEROS it can push through it's shiny metal box. THIS MATTERS WAY MORE THAN IT USED TO. Leave the hardware in the closet where the worker monkeys can't get their dirty little paws on them and run twice as many UTP lines to each work area as you THINK you will need. Wire is cheap. Except plenum wire, which is STILL less expensive than human labor.

Do it right. Do it once.

And then have your accountant amortize the hell out of it.

 

[AggieMike88]

Every bit of advice has been heard and appreciated...

Since the square footage of the space this LAN will cover is less than 3000, it will make doing a central star style setup an easier task.

My latest challenge has been finding a company to come out and do the work. I've called three so far in sequence. It has been a frustration of no return calls, "The guy who does that is out on a job and will call you later" and never does, and one onsite info gather to do an estimate who vanished into the ether with no bid to me and no return of my calls.

In the digital age with anyone having the ability to leave a Yelp or Google review about your business, providing the best customer service is key to business growth. So it's sorta baffling why the one's I've made contact are falling way short.... Or maybe not so baffling as they are demonstrating by inaction before the job the level of inaction they offer during and after the job.

I'll find someone... I just wanted to vent some of the internal pressure from the spleen.

 

[John221us]

I have done quite a bit of work with T&D Communications out of Houston. They are a good cabling company. I have worked with them nationwide, so they should be able to reach to Denton.
http://tanddcomm.com/

 

[AggieMike88]

Found someone very experienced and willing to do it.... I was looking in wrong category... Found this guy in the "Security" section on HomeAdvisor.com. Lots of positive reviews, and was at my office within a few hours of making voice contact with him.

 

[texasag93]

Another thing to watch out for is if they charge you by the run some by the cable.

Example: Contractor says it is $125 for a CAT5 run. You run 3 to one desk. They charge you $375. Contractor 2 says they charge $125 for the first and $XX for any additional wires to the same termination. Always have them terminate it to a patch panel, not just a RJ45 plug into your switch. I have done jobs both ways, it depended on what the client wanted to pay for.

Ex: http://www.frys.com/product/5958174?site=sr:SEARCH:MAIN_RSLT_PG

Cat6 cable is over 2 times the cost of Cat5e cable. With your data needs, stay with Cat5e (my opinion). This price is generally built into the wiring cost, but ask about that and see what the price difference is.

With your building, you may want to just do the runs yourself. It is very simple. 1000' box of cable and a couple $100 dollars and you will be GTG.