Network pics thread

I see nothing wrong with extra precaution here. the last datacenter I worked in, not all devices were on UPS power- Only core critical systems and network gear was. Our UPS systems only had to hold systems for 3-5 seconds for the large generator to start up, then it's all good.
 
cyr0n_k0r said:
@omegatotal and PigLover, both of you are talking out of your ass.


Really? I'd love to see your data to backup that claim.


Seriously? Any data center worth calling itself a datacenter feeds 60Hz conditioned sine wave power to the data floor. It has been fed through either generators (flywheel or other) but ideally UPS batteries with line conditioners before it even reaches your equipment.
Saying a lighting striking a data center building and it somehow bypassing the building grounds, going into the buildings electrical system, bypassing the generators and/or UPS conditioning, bypassing the 3phase 480v load centers feeding the rack rows, bypassing the rack/frame ground and going directly into your equipment is a very foolish statement.

I've had equipment in 2 different data centers in Phoenix Arizona one of the driest places in the southwest. Lightning strike alley especially during the Monsoon season. If I would ask any of the facility managers at those data centers if I should ground every single piece of my equipment because a lighting strike would make it to my switch they would laugh at me. Rightly so.
Click to expand...

Actually, no, you are talking out of your ass.

First off, the facility is not just a regular datacenter - it's an RF plant which just so happens to carry antennas that have a low-impedance path right into the back of the transmitter - which I'm willing to bet is linked in with those switches via some sort of conductive medium - be it the copper power distribution or network cable (which is also low impedance since Ethernet physical layer is very high frequency, and inductance blocks high frequencies).

In an RF plant, lightning ingress via the facility power feed is a secondary concern to lightning ingress via the antenna tower (see above note about the low-impedance (Z = 50 Ohm typical) connection to all of the equipment - although best practice in RF/broadcast engineering is to install surge suppression both at the facility power inlet and locally on devices such as transmitters (a choke around the power leads is one suggestion, by a major manufacturer of broadcast transmitters).

Rack frames do not have sufficiently low impedance to conduct the 20,000 amp to 100,000 amp surges that are produced by lightning strikes. Rack rails are not made out of copper, furthermore, they will oxidize which will increase the DC resistance. 4" wide copper strap is the preferred conductor for the grounding bus in broadcast installations (cell and microwave relays have many similar design practices). The cylindrical #6 conductors used for safety grounds have far too much inductance to be an effective ground in a lightning-vulnerable system, hence the deployment of copper buss bars running under all the racks.

Lightning is a very complex waveform that is generally approximated as a sinusoid-like series of peaks with around 10MHz of bandwidth which means that it is very much RF, rather than DC.

While an important system for electrical safety, the safety grounding system is not effective for serious lightning protection due to aforementioned inductance concerns. Furthermore, TVSS devices are not effective unless they are fitted with a low resistance, low inductance ground connection as many of them work by using MOV or spark gap devices connected to ground to conduct the surge current safely (line to line protection is also present). The PDUs would not be effective as its is not connected to a ground capable of dealing with lightning.

All of this is absolutely standard practice and facility design best practices for lightning prone environments are routinely published by major transmitter manufacturers (Nautel, Broadcast Electronics). This comes from decades worth of experience in broadcast and RF facilities, all of which have extreme susceptibility to lightning due to the large radio towers which are absolutely required for the efficacy of the facilities.
 
Since dash wants pictures... heres some damn pictures...

Finally replaced the IBM rack I used to have, downsized some so I could still have a bedroom too.



Top half of the rack- Gigabit D-link switch, works well considering I got it for 60 bucks.
The SFF desktop is my old untangle box. currently it holds up the switch because I cannot find ears for it.

The Black 4u box is my server 2008 toy box (Basically unused)
Intel Xeon 3220 2.4GHz x4
6gb ram
Evga 790i SLI FTW board
Misc Harddrives

Next is my HP DL320 G4 (Unused)
Pentium D
8gb ram
2x 80gb SATA

Catalyst 2950 series switch
Catalyst 3500XL switch
2600 series router.



The bottom half! YAY!

Top- Xyratex Fiber array (Doesnt work- Needs new controller and Im a cheapskate.)
DL385 G3 Big Vmware box Planex0
Dual opteron 275's
16gb ram
6x 72gb U320 SCSI Raid0
quad port Intel gig nic
4gb/sec Fiber HBA (Unused)

Dual Xeon Box
12gb DDR2 ram
something like 14x 500gb drives
Unused due to its power hungryness

Another older dual xeon (603's)
8gb PC2-3200
No drives in hotswap bays.
Same as top, unused due to power hungryness

Bottom is a Dual 603
dual 2.4ghz xeons, 4gb ram
no drives
unused for now.
 
cymon said:
Actually, no, you are talking out of your ass.
Click to expand...
hmm alright. Lets break your statements down.

cymon said:
First off, the facility is not just a regular datacenter - it's an RF plant which just so happens to carry antennas that have a low-impedance path right into the back of the transmitter - which I'm willing to bet is linked in with those switches via some sort of conductive medium - be it the copper power distribution or network cable (which is also low impedance since Ethernet physical layer is very high frequency, and inductance blocks high frequencies).
Click to expand...
Unless you work there any statements made in this above statement are conjecture.

haunter said:
basically Motorola made us do alot of grounding for our multi-million-dollar radio system, and we are grounding this stuff the same way. Its not that expensive especially since we already have the busbars in the floor
Click to expand...
His statement implies they are wiring this method just because they did it somewhere else.
haunter said:
its definately overkill though.
Click to expand...
Admitting the grounding doesn't specifically NEED to be there.

cymon said:
All of this is absolutely standard practice and facility design best practices for lightning prone environments are routinely published by major transmitter manufacturers (Nautel, Broadcast Electronics). This comes from decades worth of experience in broadcast and RF facilities, all of which have extreme susceptibility to lightning due to the large radio towers which are absolutely required for the efficacy of the facilities.
Click to expand...
Please let me know how any of that applies to the grounding practices of Dell switches. Did dell start making broadcast and RF equipment yesterday while I wasn't looking?

Your entire statement relies on the assumption that he is connecting outdoor equipment directly into those switches. Judging by the pictures we are shown I don't see how you can come to that conclusion. Other than the fact he is in a building with towers near by.
 
Last edited:
cymon said:
4" wide copper strap is the preferred conductor for the grounding bus in broadcast installations (cell and microwave relays have many similar design practices). The cylindrical #6 conductors used for safety grounds have far too much inductance to be an effective ground in a lightning-vulnerable system, hence the deployment of copper buss bars running under all the racks.
Click to expand...
The crummy picture makes it hard to guess, but the ground leads on those switches look a lot thinner than 6-gauge wire. I don't see any four-inch strapping.
 
cyr0n_k0r said:
hmm alright. Lets break your statements down.


Unless you work there any statements made in this above statement are conjecture.


His statement implies they are wiring this method just because they did it somewhere else.

Admitting the grounding doesn't specifically NEED to be there.


Please let me know how any of that applies to the grounding practices of Dell switches. Did dell start making broadcast and RF equipment yesterday while I wasn't looking?

Your entire statement relies on the assumption that he is connecting outdoor equipment directly into those switches. Judging by the pictures we are shown I don't see how you can come to that conclusion. Other than the fact he is in a building with towers near by.
Click to expand...

The poster specifically stated that the network infrastructure was part of the same plant as the RF equipment. If they are sharing any wiring infrastructure, RF plant grounding practices should be followed.

They would be wise to follow Motorola's practices simply because Motorola knows what they hell they are doing.

Essentially, the RF plant and tower is going to act as a current divider. The overwhelming majority of the current should flow into the ground at the tower base. Some of the current will flow through the grounded coax shields at the ingress to the facility. Hopefully by now, there's very little left to flow through the transmitter and the auxiliary equipment. However when one is dealing with currents measured in the tens of thousands of amperes then there is really no such thing as too safe, especially considering the cost of the spool of wire and time to connect the jumper versus just the TCA of a single switch.

Furthermore, the Dell gear doesn't need to have a big RF line going straight into the back - protection of auxiliary equipment in the RF plant is a very major concern. Just because it isn't wired directly into the antenna doesn't mean it's not in danger - surges can very easily flow throughout the electrical supply system, over any control or data connections, et cetera. Protection of auxiliary devices is a major concern for site planning and these dedicated ground connections are a way to address it.

@Mikeblas: the 4" strapping is used as a buss to connect devices to (and it fans out to grounding rods). Obviously connecting a 4" strap of copper to every device would be wildly impractical, so having a smaller jumper (that's probably #14 or so wire) is better than nothing. Generally actual RF hardware is going to have connections for larger conductors (although one of the problems with circular conductors is that due to something called the skin effect, very little of the actual cross-sectional area is utilized by high frequency AC - this isn't really an issue at 60Hz so cylindrical conductors work fine there. That's why straps are preferred).

There are situations in which a poorly-designed grounding system can actually make a problem worse. Since all conductors have some impedance, whenever there is current flowing through them there is a potential difference. Depending on the setup, this might cause pretty huge potential differences to exist between equipment, leading to surge-induced damage. However, since Motorola installed the system, we can expect that the ground system is well designed.

Ultimately this really comes down to the fact that when any type of RF gear is involved, the safest course of action is simply to follow grounding practices developed by the RF/Broadcast engineering industry. Yes, it's completely excessive compared to a 'normal' datacenter environment, but that is because there are other considerations - the power filtration on the utility supply may be quite elaborate but it can't do a damn thing about a strike in a separate part of the plant.
 
Last edited:
cymon said:
The poster specifically stated that the network infrastructure was part of the same plant as the RF equipment. If they are sharing any wiring infrastructure, RF plant grounding practices should be followed.
Click to expand...
Again, being in the same building doesn't mean anything if this specific equipment isn't actually connected to anything outside on the towers.

haunter said:
Basically Motorola made us do alot of grounding for our multi-million-dollar radio system, and we are grounding this stuff the same way.
Click to expand...
I don't see any radio equipment in his pictures
haunter said:
Though the move to Fiber for data to the tower has helped alot, there is still a large amount of Coax going from tower to Server Room
Click to expand...
I don't see any Coax in his pictures

Like I said before, all your statements are based on assumptions.
1) you assume he is in a dedicated RF building of some kind dealing only with RF equipment
2) you assume his equipment in the pictures is broadcast or RF equipment (which it clearly isn't)
3) that said equipment is directly connected to outdoor lightning prone equipment
 
OHH look what arrived this evening :)

Tripplite 12U rack :)

DSCN0210.JPG


No plastic :)

DSCN0211.JPG


DSCN0212.JPG


DSCN0213.JPG



Now to start stuffing this beast :)
 
Dash- Mine is better you know why? My gear is already in and I dont have to do anymore work.

You could always slip me a qnap tho :p :p
 
For those of you running servers in your own home...how do you deal with the noise? especially if you have a smaller rack with servers/switches/routers in the office and such?
 
lol, f u dash ;)

I dunno, Im used to it, have had servers in my bedroom since I was like 16, no bigs tho. mine arent super bad, but I can hear them. My switch is the worst, that fan has like a whine to it...
 
ChRoNo16 said:
lol, f u dash ;)

I dunno, Im used to it, have had servers in my bedroom since I was like 16, no bigs tho. mine arent super bad, but I can hear them. My switch is the worst, that fan has like a whine to it...
Click to expand...

so take the switch out of that SO FULL O rack and fix them :) takes ya 5 min to swap out the fans to make it silent :)
 
i dunno how to open the casing on the dlink- the ciscos are fine, dont use em cuz the copper is 100meg.
 
Also- The only reason there is open space is my 16-port KVM is on the bench behind rack, and I took out a couple empty cases.
 
techtips said:
For those of you running servers in your own home...how do you deal with the noise? especially if you have a smaller rack with servers/switches/routers in the office and such?
Click to expand...

Used a completely isolated room. Thick metal doors and walls, insulated, etc...

I run everything big in that room, and bring all drops into a patch panel. The switches and routers are bay far the loudest gear I run, so it really don't want to have them anywhere else in the house.

Works very well though, you can only hear the room if you're within 10 feet of the door. I don't have any super-loud gear anymore, like Dell SC1435 1U servers, but I would assume they would be okay in there.

I really can't imagine having serious gear not being in a sealed room. I installed one of my refurbished Fireboxes I sell in a relatives house who needed it, and their whole basement has an audible "wrrrrrrrrrrrrrrrrrrrrr" now. It's no big deal since its not really a finished basement, but rack gear is rarely able to exist outside a server room if it has fans.

Some people put rack gear right in with out-in-the-open home theater gear, I think that's crazy since you'd for sure hear it during quiet parts of a movie.
 
cyr0n_k0r said:
@omegatotal and PigLover, both of you are talking out of your ass.

Really? I'd love to see your data to backup that claim.
Click to expand...
Actually - no. I'm talking from my direct experience. I've personally seen the aftermath of severe lightning on a datacenter. Actually seen it a couple of times - both before the place was grounded properly (a disaster) and a later when it was (with almost no impact). You, however, are speaking from complete ignorance and with a form of arrogance that belies your complete lack of experience.

cyr0n_k0r said:
I've had equipment in 2 different data centers in Phoenix Arizona one of the driest places in the southwest. Lightning strike alley especially during the Monsoon season. If I would ask any of the facility managers at those data centers if I should ground every single piece of my equipment because a lighting strike would make it to my switch they would laugh at me. Rightly so.
Click to expand...
LMFAO! What a riot. Phoenix? Lightning? Really? Little kids fireworks shows are more dangerous. Spend a summer in the midwest/deep south and then post here again.
 
Last edited:
cyr0n_k0r said:
Again, being in the same building doesn't mean anything if this specific equipment isn't actually connected to anything outside on the towers.


I don't see any radio equipment in his pictures

I don't see any Coax in his pictures

Like I said before, all your statements are based on assumptions.
1) you assume he is in a dedicated RF building of some kind dealing only with RF equipment
2) you assume his equipment in the pictures is broadcast or RF equipment (which it clearly isn't)
3) that said equipment is directly connected to outdoor lightning prone equipment
Click to expand...


the COAX comes in from behind those, and then goes into the other side of the server room where all the radio racks are. there is a rack of broadcast equipment about 4 feet behind them as well

the cat6 and COAX do share cable track. and we have had servers on the network die from lightning hits even though there are only 2 network crossovers to the radio side, and one of them is fiber
cyr0n_k0r said:
hmm alright. Lets break your statements down.


Unless you work there any statements made in this above statement are conjecture.


His statement implies they are wiring this method just because they did it somewhere else.

Admitting the grounding doesn't specifically NEED to be there.


Please let me know how any of that applies to the grounding practices of Dell switches. Did dell start making broadcast and RF equipment yesterday while I wasn't looking?

Your entire statement relies on the assumption that he is connecting outdoor equipment directly into those switches. Judging by the pictures we are shown I don't see how you can come to that conclusion. Other than the fact he is in a building with towers near by.
Click to expand...


you realize that newer radio stuff is all IP based? most of all of it is Procurve switches in our building

beyond that I have a switch and a NAS in the tower for holding backups(its UPS's, generator'd, and climate controled)

cymon said:
The poster specifically stated that the network infrastructure was part of the same plant as the RF equipment. If they are sharing any wiring infrastructure, RF plant grounding practices should be followed.

They would be wise to follow Motorola's practices simply because Motorola knows what they hell they are doing.
.
Click to expand...

yeah and because we have the networks connected, motorola will audit us for warrenty/support purposes and improper/lack of grouding to their spec can void that.


dashpuppy said:
OHH look what arrived this evening :)

Tripplite 12U rack :)
No plastic :)


Now to start stuffing this beast :)
Click to expand...

niiiiiiiiiiiiiiiiiice

my home rack is such a POS
 
Last edited:
Rack%20%283%29.jpg


3x2651XM's...one will soon be running UCM if I can swing it
2x2950's
1xPIX 515E
4x HP DL380 G3's running ESX3.5 + OpenFiler

Custom built rack from an old table (for the lulz).
 
DlStreamnet said:
http://www.bitpoint.co.uk/xup/Rack%20%283%29.jpg[/img

3x2651XM's...one will soon be running UCM if I can swing it
2x2950's
1xPIX 515E
4x HP DL380 G3's running ESX3.5 + OpenFiler

Custom built rack from an old table (for the lulz).[/QUOTE]

Whats UCM?
Click to expand...
 
If it is so easy Mr Dash- I will expect your instructions on how by next week. I have the DGS-1224T.
 
^ Are there not any screws on the case? Maybe behind some stickers or sticky feet?
 
I have had all the screws out, seems almost like it was a press formed casing? like, it's folded in such a way as to not be opened again.
 
Not even a full month, from the 12th to the 29th. We've put quite a few "miles" on 3CX already.

Over 25,000 minutes inbound and almost 14,000 outbound. :eek:

1Dl3y.png


This next month should give us a good view of how the call volume will be moving forward. Our Toll Free lines are porting over, the first 3 of 11 have already done so. So there's another line item. :)

Go get you some http://www.3cx.com/
 
BEEN swamped with work these last 2 days and tomorrow is another BUSY day.

This morning i had 10 min to spare so i plopped in the rubber mounts for the fan im wiring up.

Fan will run at 7V DC and spin quiet, and no hum / vibrations :)

DSCN0215.JPG


DSCN0216.JPG


DSCN0217.JPG
 
I am sure someone does, but it is older. I have a really awesome local company that does refurb UPS of all sorts. That unit with the rails was $400.
 
Damn those are sexy. What make and model of fan and mounts?
 
dashpuppy, more info about those mounts please. I have the fans at home not properly mounted. It's a metal fan and the standard screws don't quite work.
 
You must log in or register to reply here.
Back
Top