Help - Search - Members - Calendar
Full Version: Home Network Setup
OverclockersClub Forums > Hardware > Networking & Internet
get_saif
hello one and all,

Its been a long time posted anything. here it goes, i have set up home network with 4 computers at home.

Using ( lan - hub, ADSL modem ).

NOTE NO ROUTER.

1) i have established xp on 1 comp , here after called as "A". A has enabled sharing all the harddisk to others for data exchange.

2) A shares net connection to others. ( provider)

Now i am just wondering , what else can be established with this kind of setup.

3) ?

4) ?

LINK .... http://209.85.175.104/search?q=cache:uMd_q...lient=firefox-a
.........................

And yea one more thing, as file transfer speed is 1mb/s, as the hub is a normal class made, is there a way i can increase the speed and avoid collision during file transfer.

keeptyping.gif
Syngensmyth
I'm sure it's just me, but I'm lost as to what you are asking unsure.gif unsure.gif unsure.gif
wevsspot
GS,

Any particular reason you chose not to get a router (specifically a NAT router)?

That is your first and best line of defense against unwanted intrusions into your LAN.

I'm also not sure what you're asking as far as questions #3 and #4 go.
markiemrboo
Avoid collisions, replace the hub with a switch.

Advantages of a home server? Well, my small home server does quite a lot really.

It acts as an 'internet gateway' / DIY "router" (aka good old NAT). It's a really rather flexible firewall. I have traffic shaping, though I haven't made too much use of it. These actually sort of go hand in hand - all handled with pf. I actually love this setup, and have done my best to avoid a normal boring commercial home router £50 thing smile.gif

It's a Windows domain server, and allows users to log on to it and have roaming profiles. It is a [very] small file server, so important user data can be saved to it's RAID-1 arrays. What does this mean? It means if I have to format, when I have re-joined the domain and go to log on afterwards, it's much the same! My firefox favourites remain, all my e-mails are still there (without backing up. at all) etc. Most Windows options are saved in the roaming profile, and for those that aren't I've got a log on script and policy which do a lot of it automatically anyway. Basically, I don't have to do any reconfiguring of Windows to get it how I like it after a format! It's fab.

It's a small web-server. I've designed a small simple dynamic 'intranet' for the family, which can save links, has search boxes for google, images and ebay with search history. It has a small file upload facility, which actually uploads to the site on the same machine which is accessible via the internet (the public facing site). This is per user, so it's semi-personalised. It just makes things a tiny bit easier.

I run a small caching name server. This allows me to do Windows 'hosts file' style domain blocking on a global network basis.

I run a time server. It continuously syncs it's time with multiple other time servers around the world, and my Windows machines sync to it both in the log on script, and then automatically every hour to keep clocks in sync.

I used to have a static IP (8 actually), but no longer. I can still run a dynamic DNS client, which then allows me to use SSH to remotely admin the machine, if needs be.

I do have an FTP server running accessible internally only, but that hasn't actually been used for a good while. It can be handy sometimes, and sometimes I will quickly let it pass the firewall for various tasks. Rare though.

I might even run a self coded MSN sniffer / logger, which can log any MSN messages passing through that machine (login to MSN inside my network, I get to invade yer privacy laugh.gif) rolleyes.gif I could log other things, like any sites visited for a particular machine etc.

I used to run and make use of a local IMAP / mail server. I used a script to grab new e-mail from my hotmail account and dumped them to a local server, which I could then connect to with a proper e-mail client. Thunderbird now has a plugin to get hotmail messages, so it's a bit redundant.

Most importantly, it creates my sig image with some stats about its self, and uploads it to a now rather dodgy free FTP hosting lala every hour laugh.gif



That's not using a Windows machine before you ask "how do I do...", but there should still be a fair few things you can do with some kind of central small server.
get_saif
QUOTE (markiemrboo @ May 24 2008, 09:30 AM) *
Avoid collisions, replace the hub with a switch.

Advantages of a home server? Well, my small home server does quite a lot really.

It acts as an 'internet gateway' / DIY "router" (aka good old NAT). It's a really rather flexible firewall. I have traffic shaping, though I haven't made too much use of it. These actually sort of go hand in hand - all handled with pf. I actually love this setup, and have done my best to avoid a normal boring commercial home router £50 thing smile.gif

It's a Windows domain server, and allows users to log on to it and have roaming profiles. It is a [very] small file server, so important user data can be saved to it's RAID-1 arrays. What does this mean? It means if I have to format, when I have re-joined the domain and go to log on afterwards, it's much the same! My firefox favourites remain, all my e-mails are still there (without backing up. at all) etc. Most Windows options are saved in the roaming profile, and for those that aren't I've got a log on script and policy which do a lot of it automatically anyway. Basically, I don't have to do any reconfiguring of Windows to get it how I like it after a format! It's fab.

It's a small web-server. I've designed a small simple dynamic 'intranet' for the family, which can save links, has search boxes for google, images and ebay with search history. It has a small file upload facility, which actually uploads to the site on the same machine which is accessible via the internet (the public facing site). This is per user, so it's semi-personalised. It just makes things a tiny bit easier.

I run a small caching name server. This allows me to do Windows 'hosts file' style domain blocking on a global network basis.

I run a time server. It continuously syncs it's time with multiple other time servers around the world, and my Windows machines sync to it both in the log on script, and then automatically every hour to keep clocks in sync.

I used to have a static IP (8 actually), but no longer. I can still run a dynamic DNS client, which then allows me to use SSH to remotely admin the machine, if needs be.

I do have an FTP server running accessible internally only, but that hasn't actually been used for a good while. It can be handy sometimes, and sometimes I will quickly let it pass the firewall for various tasks. Rare though.

I might even run a self coded MSN sniffer / logger, which can log any MSN messages passing through that machine (login to MSN inside my network, I get to invade yer privacy laugh.gif) rolleyes.gif I could log other things, like any sites visited for a particular machine etc.

I used to run and make use of a local IMAP / mail server. I used a script to grab new e-mail from my hotmail account and dumped them to a local server, which I could then connect to with a proper e-mail client. Thunderbird now has a plugin to get hotmail messages, so it's a bit redundant.

Most importantly, it creates my sig image with some stats about its self, and uploads it to a now rather dodgy free FTP hosting lala every hour laugh.gif



That's not using a Windows machine before you ask "how do I do...", but there should still be a fair few things you can do with some kind of central small server.


hmmmm, this is overwhelming to read through!!! Me 2 have a static ip's around.

replacing with a switch, i believe it does a decent job of avoiding collision, but still speeds which are achievable are around 6mb/s, Is that all a switch can pull in? This experiment was conducted at my frnds place....with the switch , router.

Good 2 know more about home server. But windows was the doubt, u made it clear in the end rolleyes.gif .

Me still lingering with the thoughts of adding more to the centralized setup.

thnks for showing the power of home server withstupid.gif

But would like to digg more on windows base thumbs-up.gif and wat Raid would justify for home setup, i dont want for back up or speed, SPACE is the criteria, Raid level 0,1 i think these are not budgeted. any other levels to achieve with speed and space.

thanks for the help
markiemrboo
QUOTE (get_saif @ May 24 2008, 07:09 PM) *
replacing with a switch, i believe it does a decent job of avoiding collision, but still speeds which are achievable are around 6mb/s, Is that all a switch can pull in? This experiment was conducted at my frnds place....with the switch , router.


For a 100mbit network, this is actually about typical. In theory it should be 10MB a second. In practice, you'll more likely see around 6MB to 8MB a second. If you want more than that then I guess you need to look at upgrading to 1gbit. I expect then your hard disks will be the limiting factor.

QUOTE
Good 2 know more about home server. But windows was the doubt, u made it clear in the end rolleyes.gif .


I'm sure you can still do a fair amount of all of that easily with a Windows server wink.gif

QUOTE
But would like to digg more on windows base thumbs-up.gif and wat Raid would justify for home setup, i dont want for back up or speed, SPACE is the criteria, Raid level 0,1 i think these are not budgeted. any other levels to achieve with speed and space.


RAID for space? Then you definitely don't want RAID-1.

RAID-0 - Speed and space. You'll get speed improvements and won't lose any space, because there's no kind of redundancy going on. Not for critical important data really! If a disk in the RAID-0 array fails, you're buggered basically, as data is striped (spread) across all drives in the array. This is where it gets it's performance from.

RAID-1 - You may actually find you get *read* speed increases, especially in a multiuser environment. This may depend on the RAID implementation, but as data is mirrored over multiple disks it means that in theory one process can read whatever data it needs from one of the disks, while another process can read the data it needs from the other disk.... at the same time (if the disks are on separate controller channels in the case of IDE)! Writes may be a tiny bit slower, as the data is duplicated identically across each disk in the array ("mirroring"). This means that for 2 x 80GB drives in a RAID-1, you get only 80GB of usable storage space. However, if one disk fails you still have your data mirrored on the other drive. I use two separate RAID-1 arrays, one for the O/S and one for user data. The O/S can even boot from either of the drives in it's array, which is quite neat.

RAID-5 - You could probably consider this as being in between 0 and 1. You get redundancy, more space available than RAID-1, but less than RAID-0. You'll need three drives minimum. It stripes the data across all drives, and also adds parity information across them all. When you add up all the parity data, it will actually consume the entire space of one of the drives (though it's spread across all of them). This means that for, say, 4x1TB drives you should get 3TB of usable space, where as RAID-1 will only manage 2TB and RAID-0 will be able to use the full 4GB (but again, with NO redundancy and possibly increased risk). You should get *read* speeds close to RAID-0, but write speeds may not see much of an improvement because of overheads. Due to parity calculations required, it is generally preferred to get a proper hardware RAID controller for RAID-5, rather than using a software implementation which will use your CPU for these calculations and slow things down.

If you want pure space, no redundancy, but don't want the risk associated with RAID-0 then there is JBOD. This can 'combine' any sized disks together to create a single larger volume. The data isn't striped across all drives like RAID-0, so you probably wont really gain much in terms of speed. *However* a single file may still actually end up spread over multiple disks due to the way file systems work. Regular defrag may help to keep large sequential files close together and on the same physical disk in this case, but even then still doesn't guarantee that the entire file will be on one physical disk.
get_saif
QUOTE (markiemrboo @ May 24 2008, 11:48 AM) *
For a 100mbit network, this is actually about typical. In theory it should be 10MB a second. In practice, you'll more likely see around 6MB to 8MB a second. If you want more than that then I guess you need to look at upgrading to 1gbit. I expect then your hard disks will be the limiting factor.



I'm sure you can still do a fair amount of all of that easily with a Windows server wink.gif



RAID for space? Then you definitely don't want RAID-1.

RAID-0 - Speed and space. You'll get speed improvements and won't lose any space, because there's no kind of redundancy going on. Not for critical important data really! If a disk in the RAID-0 array fails, you're buggered basically, as data is striped (spread) across all drives in the array. This is where it gets it's performance from.

RAID-1 - You may actually find you get *read* speed increases, especially in a multiuser environment. This may depend on the RAID implementation, but as data is mirrored over multiple disks it means that in theory one process can read whatever data it needs from one of the disks, while another process can read the data it needs from the other disk.... at the same time (if the disks are on separate controller channels in the case of IDE)! Writes may be a tiny bit slower, as the data is duplicated identically across each disk in the array ("mirroring"). This means that for 2 x 80GB drives in a RAID-1, you get only 80GB of usable storage space. However, if one disk fails you still have your data mirrored on the other drive. I use two separate RAID-1 arrays, one for the O/S and one for user data. The O/S can even boot from either of the drives in it's array, which is quite neat.

RAID-5 - You could probably consider this as being in between 0 and 1. You get redundancy, more space available than RAID-1, but less than RAID-0. You'll need three drives minimum. It stripes the data across all drives, and also adds parity information across them all. When you add up all the parity data, it will actually consume the entire space of one of the drives (though it's spread across all of them). This means that for, say, 4x1TB drives you should get 3TB of usable space, where as RAID-1 will only manage 2TB and RAID-0 will be able to use the full 4GB (but again, with NO redundancy and possibly increased risk). You should get *read* speeds close to RAID-0, but write speeds may not see much of an improvement because of overheads. Due to parity calculations required, it is generally preferred to get a proper hardware RAID controller for RAID-5, rather than using a software implementation which will use your CPU for these calculations and slow things down.

If you want pure space, no redundancy, but don't want the risk associated with RAID-0 then there is JBOD. This can 'combine' any sized disks together to create a single larger volume. The data isn't striped across all drives like RAID-0, so you probably wont really gain much in terms of speed. *However* a single file may still actually end up spread over multiple disks due to the way file systems work. Regular defrag may help to keep large sequential files close together and on the same physical disk in this case, but even then still doesn't guarantee that the entire file will be on one physical disk.


Good info, reminds me of my 2nd sem BCA days:), will try levels and report for furthur assistance,.......

By chance if you know a good switch , which can be good in all areas, please LINK it. Will look into it also simultaneously.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.