Benefits of buying a "real" server?

Hi all,

When we started our business two years ago, I built a mid-to-high-spec desktop machine with Windows SBS 2008. It works well.

I'm ready for a new server: forgetting my individual requirements, why should I buy a "real" server? For the price of some of the HPs, I could build something much more "powerful". I believe "real" servers are more reliable - are there better performance outputs from the seemingly lower specs?

Thanks in advance!
Who is Participating?
Generally server have better quality parts, faster internal buses and redundancy - 2 power supplies, hardware RAID and they will usually have longer guaranteed support and replacements - desktops change so often, replacement parts can be an issue over time

Many of the "server" machines, have redundant power supplies, built in disk RAID controlers, and much better cooling than a desktop, redundant and monitored fans etc.  As well as the usual multiple LAN ports etc.  Many can also support hot swapping of drives, and some other sub-systems, that makes maintenance a little easier, as you could do things like air filter cleaning without powering the thing down.

If you go for one with dual PSU's, then make sure you run them from different circuits, fed by different circuit breakers etc, but adhering to local electrical safety codes etc.

If a dual PSU server, I guess idealy both should have a UPS each, but so long as one of them is fed via a UPS that "should" be good enough for most needs, so long as the Server can talk to and monitor the UPS(s) so as to do an orderly shutdown in the event all the lights go out for an extended time, when you're not there!

The multiple LAN thing, is so you can segregate various internal and external networks, controling access between them by rules on the server.   You wouldnt want for example, the accounts dept to access the production floor PC's, or the other way round, but they could both need access to common areas on the server for other purposes, as well as backups etc.

It takes a lot of thought and planning, to get that sort of thing right.

Have Fun!.

A real server should come with fault tolerance and redundancy galore.  I.e. RAID1 or RAID 5 or RAID10.  This means if a single disk fails your' not toast. They have redundancy drive controllers, redundant power supplies, server class hard driver, and many other features.  Many of the drives are hot swappable. One dies, through another in the RAID without even powering off.  The dell PowerEdge series are pretty nice. You can pick up an old 2950 pretty cheap and they do nicely. It depends upon your needs.  Oh yeah, they usually have dual Gb ethernet cards as well and features to get to the machine even if the OS is hosed.
Ultimate Tool Kit for Technology Solution Provider

Broken down into practical pointers and step-by-step instructions, the IT Service Excellence Tool Kit delivers expert advice for technology solution providers. Get your free copy now.

PowerEdgeTechIT ConsultantCommented:
As has been said, REDUNDANCY is one of the key reasons to get a server, as it protects you from failing power supplies, failing disks (although a backup strategy is still needed), and even failing controllers, network ports, and memory (depending on the configuration and system capabilities).  They also typically hold much more memory and up to two or four processors for serious applications and multiprocessing.

If I build the machine for some significant function, then it crashes and burns horribly, that's all my fault.
If I buy a whatever brandname server, which then crashes and burns horribly, this is brand vendor's fault.

Same principle applies to mismatches specs for the machine.

So, buying a "real" server allows to shift some responsibility to server's vendor, because they are supposed to be wise and reliable. Should something fail, it is partly their fault.
PowerEdgeTechIT ConsultantCommented:
While I don't like the idea of shifting the blame (did you then choose a bad vendor?), the benefit here is that you have someone to address the problem with, on behalf of the customer ... you have a partner - a team - to assist you in getting your customer's problem taken care of.  You call Dell server support or HP server support when there is a problem, and they will assist you, the company to repair or return things to normal.  If you call support for a workstation and explain to them that you are running all the company's data software on a workstation, the first question will be "why are you running mission critical data and services on a workstation and not a server", and they are simply not equipped to deal with those kinds of situations.  Server support is geared toward and trained to deal with "mission critical" situations and will be better able to 1. understand your problem and urgency, and 2. help you get it squared away.
I agree with PowerEdgeTech. And it isn't going to matter whose fault it is if all "hits the fan." You will be required to get it going again ASAP.

A good backup strategy is a must. Preferably a regular data backup and a BMR (bare metal restore).
If you need continual backup protection, R1Soft has a great product for this purpose. It can do BMR and can be set to backup changes every X minutes, so you could get a shot every 15 minutes of the day if need be.

Also as mentioned if through Dell, and you  have their support you can get help.  You can also get drivers and software easily through them via their machine specific identifiers.

I would further say if it is important and you can afford it, buy two identical servers and leave one disconnected because lightning won't care about the redundancy.

There is also form factor. If you have a rack you can get nice 1U, 3U, etc. servers that fit nicely on rails and in the rack.  

I would also recommend a good UPS from someone like APC or MinuteMan.
years ago I had the same discussion with some one.  They had an old compaq and new hand built server for a small business.  The Friday evening before a 3 day weekend a storm took out the ac unit on the roof.  The compaq turned itseful off after reaching a certain temp, the hand built unit fried the drives.    He learned his lesson.  But you can other software monitor the enviroment and shutdown the equipment.   The hardware, like everyone has mentioned, is made to run 24x7.  
Warren_RAuthor Commented:
Thanks everyone.

The 24x7 is the thing that sticks in my mind.
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.