Does having multiple RJ45 jacks degrade the Internet signal a lot?

I've had really bad luck with Netgear, which is why I suggested Dlink. YMMV. Dlink has metal gate versions of their gear.

Regarding double NAT, that is really annoying. Some routers can detect the double NAT (don't ask me how) and warn you.

Most DSL modems are one port routers. This can lead to address conflicts. I really wish the modem manufacturers just expected the customer to use a router. I don't know one person with broadband that doesn't have a router attached. AT$T stared selling DSL modems with routers probably to stop the customer service calls.

That was the original plan.

Nope.

I only need ONE in the game room.

What you are saying hits on the original design objective. I wish my command of English were better because I confused everyone (I think) because I'm confused HOW to wire from the router to the game room in this scenario.

Is this the correct scenario for the sum total of the house wiring?

1. Cat5 cable enters house at lowered garage at the near end of the house
2. Cat5 continues into crawl space & into office floor in house center
3. Cat5 ends at a single female RJ45 in the (centrally located) office
4. Another cat5 cable starts at another single RJ45 in the office wall
5. That (inactive) cat5 goes down into the crawl space to the game room
6. That (inactive) cat5 ends at a single RJ45 jack in the game room

If that is the correct wiring sequence, then I can do that relatively easily.

Now comes the active connections.

In the office - this is what I was planning: a) The POE sits in the office, one end connected to the wall plate. b) The other end of the POE connects to the Linksys WRT54G broadband wireless router (which also has four LAN ports in the back) c) One LAN port of the WRT54G goes to the desktop computer d) Another LAN port of the WRT54G goes to the Belkin VOIP desktop phone

At this point, everything but the game room is now working. The problem is that the game room is too far away for a good signal out of the WRT54G.

I have two options (I think) for the 'game room' at the far end of the house: I. Add a wireless "repeater" of some sort (the purchase & setup of which I am unfamiliar) II. Add a cabled connection

THIS IS THE PART THAT WAS CONFUSING ME IN THE BEGINNING:

If I go with the cabled connection to the game room, is it 'this' simple? A. I attach a jumper from the LAN port on the back of the WRT54G to the second office wall plate ...(this 2nd office wall plate is just a connection to the game room wall plate through the crawl space) B. I attach a jumper from the game room wall plate to the Wii

My original (I agree confused) question was: Is it 'that' simple to add the game room as a wired connection?

NOTE: It seems weird to me to have a 'dead' wire simply going from the office wall plate to the game room wall plate.

The salmon colored Home Depot outdoor-rated solid conductor cat5 cable ($75 for 500 feet) is 24 AWG.

I'm guessing it's ~100 feet to the office in the center of the house (including zig zags inherent in routing the wire).

The POE is 15 volts (but I have no problem buying a higher-voltage POE if that is what is needed).

I guess I 'could' go up in size to 23 AWG cat5 cable - but the initial comments intimated this gauge solid wire should be find for three times the distance I'm calculating. (I think.)

Jeff,

Will this device work with ethernet that includes Power? My wisp installation uses POE from inside my computer room out to the radio on my antenna mast. I don't know but think it's just plain ethernet from there to my router. then plain ehternet after the router.

In California, all the cities and counties have low voltage wiring inspections as part of code compliance. In most cases, they simply require compliance to the latest NEC wiring codes. They may add their own details, but the basic requirements will need to be met. Incidentally, the Peoples Republic of Santa Cruz requires a permit for any construction costing over $500.

It's not the wiring that's the problem. It's how it's mounted and what it's made from. For example, you need to run plenum cable through air spaces. Plenum cable does not generate much smoke and will therefore not asphixiate fire fighters. Proper support and using riser cable for long vertical runs is simply best practices to prevent the wire falling or breaking under its own weight.

Try this quiz for practice:

Plenty more:

From my limited and somewhat dated experience, the inspector doesn't care much about sloppy LAN wiring. He's probably a former electrician or contractor and doesn't know much about LAN wiring anyway. He does care that the low voltage wiring is at least 2" away from AC power, that it doesn't share any wall outlet boxes, and that it's not running high currents through the cabling. Incidentally, some IEEE spec recommends 6" for 120vac and 12" for 240vac. Where there are few cables, the inspection is rather superficial. Where there's a large number of cables (hospital, corporate, data center, etc), the inspections are more thorough.

Thank you for that advice (as I was parroting the word "active" from advice given earlier in this helpful thread).

I do greatly appreciate this clarification (just as I did in 'wire' versus 'cable').

I'm already confusing enough to try to understand so I'll try to use 'wire' as a verb; and to drop the 'active' adjective in ethernet switch.

You can ALWAYS correct me as I appreciate the subtleties!

I bought several of those switches. Gigabit works well enough, but I was never able to get close to wire speed with them. I'm not sure if it's the switch or something else. Caveat Emptor.

Huh? An ethernet switch works on ISO layer 2 (MAC layer). DHCP works on layers 2 and 3 (IP layer). As long as the switch can pass broadcast packets (they all should), you should not have any problems with DHCP broadcasts and negotiations. Each port on the switch has its own MAC address. Your DHCP server should be picking up the MAC address of the originating computer, not the local switch. If it grabs the switch MAC address, then yes, it will try to change IP address every time you move the ethernet port. However, that's NOT the way it should work. Double checking:

C:\>arp -a Interface: 192.168.1.11 --- 0x4 Internet Address Physical Address Type 192.168.1.1 00-16-01-97-fd-a6 dynamic

Yep... that's the MAC address of my Buffalo WHR-HP-G54 wireless router, and not the MAC address of the ethernet switch located between my PC and the router.

No comment. I won't generalize by manufacturer. Each one has their winners and their losers. Dlink seems about average.

I recently picked up several Dlink DIR-601 (N150) wireless routers.

These are cheap and basic routers. So far, no problems or failures. I expected problems due to the new "green" features, such as reducing the ethernet transceiver power for short cable lengths, but so far, so good.

Hell no! You're using it wrong. It's only intended to check wire continuity. That means you have to unplug BOTH ends of the cable run. Plug one RJ45 end into the "transmitter", and plug the other RJ45 end into the "receiver". Then, you watch the light show at the receiver end, making sure that there are no missing or out of order lights.

Note that the cable is completely disconnected at both ends from the PoE injector and splitter. These won't interfere if since they're not connected to the tester.

Unplug the cable and test just the cable.

Not totally true. There were ethernet hubs. You likely can't buy one any more.

from "state of the art netwoking":

By default hubs are single broadcast and single collision domain, which means a device transmitting at a time, transmits to all the devices in the network i.e. it broadcasts every time and every device on the network listens to that broadcast and the one which it is meant for picks it up. It?s anyone?s guess that how efficiently it will work, its okay with one or two or three devices in a network but with network scaling up and more and more devices being connected to it the network dies down. How often we listen the complaints in office or home that the network being slow or down, if there are hubs in the picture that?s what going to happen, because there is no way with hubs you can control LAN traffic congestion. One way to make an ever increasing network is to segment a network in smaller part and that?s when the switches come into picture.

Switches are much more than multi-port repeaters, they are quite intelligent in a way that they recognize the devices connected to it by their addresses, so there is no need to broadcast every time one device want to share something or exchange information with another device. It?s like now when hubs are gone I can talk to my friend by addressing him by name, otherwise with hubs it was like I had to shout from the rooftop for everybody to listen even though they didn?t want to, what I wanted to say to my friend. So the above explanations make switches a single broadcast and multiple collision domains. It broadcasts only in one scenario in which it does not have information about a device in its mapping table for which a particular piece of info is transmitted, so it broadcasts that info that one time and after finding about the device which accepts that it updates it table.

Also hubs operate in half duplex while switches can operate in full duplex mode too. Adding a switch adds a lot of functionality to the network and improves the efficiency of the network too. You can still use hubs as per your networking needs but try using at least one switch in case of a multiple hub network by plugging the hubs to the switch, but an all switched LAN is just always better and I think I?ve provided enough evidence for that.

Jeff's words are very interesting words of experience.

It's exactly what I don't have ... so I doubly appreciate the advice!

Especially since I'm finding walls within walls the more I drill deep! :)

Here are some pictures of the setup (to explain what I mean).

Picture taken just now of the WISP antenna setup (jury rigged with extension cords and patch cords until I get the wiring figured out).

Here is what I found when I popped a hole in the game room wall! (there was a hidden wall inside the outside wall!)

I can certainly believe the common box issue. Well separated boxes would never pass muster with the lady of the house. Bad enough they try to block outlets with furniture. Your suggestion of wiring two walls on either side of the door is a good one. Almost any wire in a room can be tolerated except if it crosses a door.

Years ago I took a structured wiring "class" at CES, just to see what was happening. This was before the WWW was cranking at 11, though it existed. They suggested two networks per room. I could never get a reasonable explanation for why this was a good idea. Not on different walls, but two networks to the same outlet. Like the person saw it done, but didn't really know why himself. Of course there is no shortage of space on the wall outlet plate for multiple RJ45.

That was where I learned nearly everyone in the class was an ex-alarm installer. Oy!

If and when the "puppy attack" happens you cut the damaged cable and install a jack. Always leave some extra cable in the wall or wherever so you have something to work with.

Yes, it's that simple. That's exactly how I've done it myself.

It's not dead once you connect devices to each end of it.

Being in the business, I have had more D-Link consumer equipment failures than any other brand. Their commercial grade stuff is so-so. The only one's I've used have failed within 5 years. I'm currently using an Airlink NIMO router in the basement of my 2 storey house and the signal is useable throught the whole house and 15 feet behind the house on the patio as well.

That doesn't make sense. Why would a single port router lead to address conflicts?

If the DSL modem already includes a router, why are people adding a second router? I know there are some edge cases, but I'm wondering about the majority.

I'm just running the switches in rooms (short cables), but a friend has them going through the house, much like what the OP wants. I don't know if he ever checked them for speed.

I might just get a Dlink router and chalk the Linksys box (not a cheap one) up to experience. The one time I needed support from D-Link, it was excellent. I can't say my experience with Linksys is the same.

I fired up some Dlink WAPs about two years ago. Old B stuff I no longer used, but wanted to do a quick bridge. They still worked. I've had Netgear stuff fail under two years. Twice. However, they may not suck anymore. Or maybe I got junk. People complain about Netgear power supplies, but mine were fine. I saved the wall warts and trashed the rest.

Netgear used to have a different name. It was a product of a merger IIRC>

Ah yes, Bay Networks.

According to wiki (which of course doesn't mean much), Netgear is outsourced and D-Link is not. I don't get too bent out of shape regarding OEMs, but I ODM-ing is another story. Support on ODM gear tends to be poor. When you use ODMs, then you are just shipping black boxes.

I'm just appalled at the crappy service I got from Linksys on a box close to $200.

Here, to help, are 6 pictures uploaded to simplify descriptions:

1. WISP antenna currently jury rigged with patch cords & extension cables:
   formatting link
2. So-called "bucket router" temporarily being used to bring signal into the house wirelessly from outside:
   formatting link
3. Crawl space view up into the floor of the office (center of house):
   formatting link
4. Crawl space view over to the game room (far end of house):
   formatting link
5. Entry point at the office (where the office jack will be placed): www1.picturepush.com/photo/a/7212874/1024/Anonymous/cat5-questions.gif
6. Entry point of the game room (where the game room jack will be placed):
   formatting link

Nope. There were no *PASSIVE* ethernet hubs. Passive means that there's no powered electronics inside. You can combine (mix) data on a token passing network, such are Arcnet, but not with ethernet. You could consider 10base5 or 10base2 to be some manner of distributed passive hub, but that's not common terminology.

If it's in print, it's obsolete.

Correct. A hub is also called a multiport repeater (especially in IEEE documents, which drive me nuts).

Detail:

1. A two port ethernet switch is called a bridge.
2. All 802.11 wireless is bridging. While there may be layer 3 IP configuration for the router section, the actual wireless traffic is bridging.
4. >It?s anyone?s guess that how efficiently it will

I haven't seen many large hubs for perhaps 15 years. Compex TX3264U

64 port hub is one. Cisco had a large hubs, but I can't find the number. They not common.

What limits the speed of the hub is that it can go no faster than the rated wire speed of a single port. For example, if I had a 100baseT hub, and was running a network backup between two ports, there would be zero bandwidth available to the other ports. Needless to say, hubs don't scale very well and are easily maxed out.

Ethernet switches don't have that problem. There are two basic types, bus and crossbar. The bus type bandwidth is limited by the bandwidth of the internal backplane. 2GHz is typical. You can transfer data between any two ports at wire speed. You can also have independent (non-blocking) transfers between two other ports, up until the available bus bandwidth is exhausted. For example, a 100baseT switch, with a 2GHz bus, can use up to: 2000 / 100 = 20 pairs of ports. That limits this particular switch to 40 ports before it runs out of bus bandwidth. The crossbar switch is simply a cross point switch between any two ports. This becomes unwieldy with a large number of ports, because every time you double the number of ports, you need 4 times the number of cross points. It is cheap and effective up to about 32 ports. Of course, there are hybrids between the two types.

That's called collision domains.

Not quite. Switches pass broadcast packets to all ports. This is one of the big headaches with large wireless LANs, which can easily (and quite often) end up belching nothing but broadcasts (ARP requests etc). I've sniffed a (now defunct) muni wi-fi system that was doing that. Little wonder users found it slow and useless.

What ethernet switches will NOT pass are collisions, corrupted packets, malformed packets, garbage, jabber, and noise. Not passing collisions is why it's called a collision domain.

Discarding packets is something that an ethernet card, in a PC, does quite easily and neatly. The card just looks at the header and determines if the destination is the local device. If not, it just drops the packet. This is all done in hardware on the ethernet card and does not involve the CPU. The time wasted by dropping packets cannot be recovered, which will certainly have an impact on speed, but it will not slow down the computah.

Yep.

Sorta. Many managed ethernet switches (i.e. Cisco) allow some ports to be configured as a hub. These are intended for monitoring traffic on other ports. Since the monitor port will not be used for any incoming traffic, there's no slow down caused by it monopolizing the switch bus bandwidth. It's very handy for network management, traffic monitoring, diagnostics, snooping, and playing around. Otherwise, hubs are a bad nightmare that are thankfully obsolete. A clue is that you can't buy a brand new ethernet hub anywhere.

Been there. Use a recessed wall plate:

I dunno about this one:

Yep. It didn't take me much to work out the problem and magic formula. Convincing the customer is much more difficult.

I've never heard that one.

I've had an electrician tell me that the right way to wire a house is to put run 4 wires instead of 3 to each wall outlet, and set them up so that each of the two outlets goes to a seperate breaker, and possibly a seperate phase. It's also handy for wiring 220VAC in the same outlet. The extra wire can also be used for wiring 3 way switches. Probably a good idea, but nobody is doing it.

These daze, home alarms are either carrier current on the AC power wires, or wireless.