network performance

[Merlin]

I didn't know the answer but I am sure one of you gents/ladies do. In a business computer network that uses Windows XP pro as the base system with a custome inventory/invoicing program, how much does the hardware of each indevidual PC affect the speed of the whole network. At work we have 7 servers serving about 250 workstations, I looked and in my office there are 2 PCs with 1.4 Ghtz processors and 128meg of ram and one 2.4 ghtz with 248 meg of ram. Would this slow down our network? Would it slow down IF there were say half of the PC that were SLOW and half fast? Please explain your answer IF you don't mind. Oh and IF someone tries to browse the internet while running our custome stuff....serious bog

[Darkstar]

by slow down the network, do you mean access to file servers? Internet? What is your wiring, what type of switches and what kind of servers all have an influence on the networks performance also.

[Merlin]

primarily I mean what I think is refered to as file server access. To be sure what I mean I will explain what we do. We are a sales company so our "system" is an inventory maintenance, invoicing, accounts recievable etc. for 80+ locations in 5 states. By slowing down the network I mean while I am entering an invoice into our computer it will sometimes take a minutes or two for a line item on the invoice to "finalize". I know all the other factors that could influece the speed here but I do not know IF the hardware configuration of each indevidual terminal plays a role or not and IF it does how large a role does it play. To simplify.....IF the company had up to date terminals rather than old outdated pieces of......would the system "lag" less??

[kenc51]

primarily I mean what I think is refered to as file server access. To be sure what I mean I will explain what we do. We are a sales company so our "system" is an inventory maintenance, invoicing, accounts recievable etc. for 80+ locations in 5 states. By slowing down the network I mean while I am entering an invoice into our computer it will sometimes take a minutes or two for a line item on the invoice to "finalize". I know all the other factors that could influece the speed here but I do not know IF the hardware configuration of each indevidual terminal plays a role or not and IF it does how large a role does it play. To simplify.....IF the company had up to date terminals rather than old outdated pieces of......would the system "lag" less??

Assign a static IP for the machine doing the invoices!

[bubba]

I don't think the old PCs are what is slowing the network down, but rather slowing the server down.

What it sounds like is that you server is waiting for the slow client PC to respond that is its ready for the next step. If the server will not do anything else until it has received the 'I'm done, whats next' signal from the client then that is where the lag or slow down on the server is coming from.

If this is the case then a PC that will run the client software faster will then in turn speed up server response. Look it at like this, talk to some one who stutters and someone who doesn't, which one will finish the conversation first.

[Merlin]

The stuttering PC answers my question. I was mostly wanting confirmation that my "bosses" are not too bright. I my office I have 3 PCs 2 are 1.4 ghtz with 128 of ram while one is newer with a 2.4 ghtz and 248 of ram...in this situation don't the two 1.4s make the 2.4 as slow as them...a chain is as weak as its weakest link??

[DL126]

The stuttering PC answers my question. I was mostly wanting confirmation that my "bosses" are not too bright. I my office I have 3 PCs 2 are 1.4 ghtz with 128 of ram while one is newer with a 2.4 ghtz and 248 of ram...in this situation don't the two 1.4s make the 2.4 as slow as them...a chain is as weak as its weakest link??

There are dozens of determining factors, therefore it will be almost impossible to give u an accurate answer here.

But the short answer would be no.
The hardware of an individual workstation has no influence on the operational speed of any other workstation on a network.
The server you're working off of would more likely be the culprit IF you're running a server side app and not a simple peer to peer file sharing operation.

And there are no fewer than 50 more of those IF's involved so I hereby refer back to my opening statement.

[Merlin]

I am not being argumentative but rather want to learn. If what you say is true why do gaming servers advertise they will "boot" anyone with a high "ping"??

[DL126]

A high ping time has more to do with one's connection speed to the game server & the bottlenecks between him and the server than individual machine capabilities.

If you had a P.O.S. computer hooked to the Internet backbone, you'd have a low ping time.
But you'd still get your a$$ blown away in a FPS simply because everyone else's machines are faster at processing what comes over the connections.
Those gamers you mentioned would enjoy someone coming into a game with a low ping and a slow PC.
They LOVE target practice and a chance to improve their "batting average."

Example:
The wife and I just finished a LAN game of UT3.
We're on a GB network.
But what I see on my screen is much better and faster than what she see on hers.
Has nothing to do with the network, only what's presented on screen by the hardware within the PC.
And it doesn't slow my PC down to match hers.

Clear as mud?

Didn't take it as being argumentative.
Have a beer on me ....

[Merlin]

I had a beer...well I had several thx. Now to make sure I understand. All our inventory data is stored on the server(s) and as the business day goes on and each location is "pulling" data from those servers and modifying the numbers based on what is sold to whom where for how much, the speed of the indevidula PCs accessing the server does not affect the speed of other PCs accessing the same data?

[DMB2000uk]

I don't have a clue how it really works, but based on that description, it would make sense that when the clients pull the data from the server to edit it, it is locked on the server until the client updates it and flags it as finished.

If this is the case then you will find that you will be waiting longer (than if all the machines were fast) for the records to become available from the slow machines, but the faster machine will still be able to process/edit its data as fast as it can, it will just have to be held up when waiting for the other clients to update to the server.

I've probably over simplified what goes on there, but it echo's what bubba said, and also in a way DL126's comments.

Hope it made it clearer for you.

Dan

[DL126]

... All our inventory data is stored on the server(s) and as the business day goes on and each location is "pulling" data from those servers and modifying the numbers based on what is sold to whom where for how much, the speed of the indevidula PCs accessing the server does not affect the speed of other PCs accessing the same data?

Nope. IF its a server side app.
In that case, the pc's are nothing more than "dumb" terminals. The server(s) are responsible for doing all the processing.

Are all of these servers in the same building?
And all remote workstations are coming into that same building?
That very well could account for a network slowdown. Having a direct effect on the slow web browsing you mentioned.

As I mentioned in my first post, there are so many variables it's impossible for me and my feeble brain to make a "remote" diagnosis.

[Merlin]

I know that there are a hundred things that influence how fast a workstation runs when linked to a network and I know that there are a few things relating to the hardware of a workstation that ONLY affect that workstation. I also know that the servers themselves are the largest factor in the "speed equation" . I guess what I am asking is for an explanation as to how a multiple server multiple workstation network actually works. I am not good at graphics in a word program so I will try to explain what our network is made up of and more detail about what we do with it. This is all heresay but we have 7 servers in seven different locations all of which are "supposed" to be mirrors of each other. We have about 250 workstations in about 80 different locations spread accross 5 states. I say "about" because I do not know the exact numbers. We are a supply company so in each location there are people selling merchandise to customers all day everyday. Each location can look at the inventory of another location to see if they have an item available to sell they can also look at the sales history of a given customer to see what that person purchased for how much and where. As far as I can tell the inventory numbers are updated live so that IF someone were to look in my stores inventory as I was selling something they would see the number available drop as I keyed the invoice. They would also see the amount a customer owed on his open charge account drop by the amount of the payment he just made within seconds of it happening. I can also print out a list of everything a given customer purchased last year IF I needed to. In this setting is the hardware that makes up a workstation ONE of the factors that Influence workstations speed? if it does to what degree scale of 1 to 10 ten being the most and IF not please help me understand why it wouldn't.

[Pesto]

I would say that the network switch may be contributing to the delays. Also if this is an old server running multiple applications then it may need a hardware update. some of the problems could also stem from the configuration of the server, the disk array has to access data for several users apparently so the raid configuration could slightly contribute to the delay for users, so if its running a Raid-5 it would access data quicker than say a Raid-1 for hundreds of users. Next thing is if you're accessing a remote server over a WAN. that could be a router slow-down or a lack of throughput (like using DSL rather than a Frame-Relay connection)... either way theres a butt-load of stuff to look at and usually calls for an on-site consultant to evaluate and then sell up some fancy new equipment

and im sure theres things im overlooking, cause im not getting paid and dont really know the environment.

[Merlin]

... All our inventory data is stored on the server(s) and as the business day goes on and each location is "pulling" data from those servers and modifying the numbers based on what is sold to whom where for how much, the speed of the indevidula PCs accessing the server does not affect the speed of other PCs accessing the same data?

Nope. IF its a server side app.
In that case, the pc's are nothing more than "dumb" terminals. The server(s) are responsible for doing all the processing.

Are all of these servers in the same building?
And all remote workstations are coming into that same building?
That very well could account for a network slowdown. Having a direct effect on the slow web browsing you mentioned.

As I mentioned in my first post, there are so many variables it's impossible for me and my feeble brain to make a "remote" diagnosis.

I know my question was answered here I just don't understand why it is. I sort of look at it as if there were 5 people reading the same book and they all read at different speeds...how can it be that the slowest reader doesnt set the pace??? Like I said I know I have recieved my answer I just dont understand how it could be true.
I didnt ask for evrything that affects a worstations speed I asked IF the w0orkstations speed can affect the network and IF so to what degree...

[DL126]

I'm sorry I, I simply don't know how to explain it any better.
I'm not a good teacher.

IMHO, the the degree to which one workstation can affect another on a network is 0.

If you took the slowest workstation out, and replaced it with the baddest thing out there, it would still be faster than the remaining carry overs in local terminal operations.
During server side operations, that shiny new bad a$$ would be no faster than the server can deliver data to it.

In your example, the reader reading aloud would in fact set the pace for the rest, thereby representing the server in a network. NOT a workstation.
If those 5 people were reading to themselves, they'd still be reading the same book, but the fastest reader would get finished first, having no effect on the other 4 readers.

And lastly, to borrow a sentence from Pesto .... "... cause im not getting paid and dont really know the environment"


Best I can do for you.