Subject 1:
@Ches I would not use the guest feature of a router in a business environment. The guest feature is for home use. The key is you do not want the Wi-Fi on the same network. With the guest feature, while they are controlled somewhat with a password and some restrictions, they are a small hack away from your data. If you have static IP addresses and multiple ports in your cable modem, you can use a different static IP and a different subnet (your regular network is 192.168.1.x, then you would want it on 192.168.16.x or whatever different third octet). This essentially makes it like it is on an entirely different network down the street. If your modem has one port, run it into a $45 switch BEHIND your main router. Then run patch cords to your main router and to your wireless Wi-Fi. When you configure your wireless, use a different static IP as your public address, use the same default gateway and likely a 255.255.255.248 subnet (depends on your ISP).
I know nothing of the fonera box. Sounds cool. I don't know if it is isolated. I will trust Ken. But, if it uses the same subnet, I would be wary. I am surprise people don't use it more. Where I am, we don't use wireless, ironically we only do for our patients, but the next door office does, and people try to get there all the time.
@Tom First, let me know where I can send you a case of beer. FINALLY!...the first person to ever have acknowledged that there are other programs than AC on a computer. You have truly made my day. Now, point by point:
1. This proves the point. What are crashes. We all can have them, even with the best of servers, but crashes are unacceptable. They should not be accepted as part of having inexpensive hardware. How could you improve with server/client. Once again, are we walking server/client or one with a domain. Think of a Ferrari. Certainly better than a Corolla, but if you can't take it out for a 100 mph spin, then it gives you nothing. If you have a domain, but you don't know how to leverage GPO or Active Directory, it won't help much at all. Backups are better to some degree. Networking in general is better. I apologize if I made server class hardware out to be that much better, but it is much more stable. You can share folders on workstations, but you can share them more easily on a domain. Not many on here run two drives in RAID configs with all of their shared folders on the second drive isolated from the rest of the server. You share that one partition and everything is shared, then you can put permissions at a granular level or vice versa. Don't get me wrong, this can be done on a workstation as well, but not many workstation motherboards will allow you 20 to 64 GBs of RAM.
2. Don't purchase that computer. I would get a Dell or HP and I would buy DIRECTLY from the manufacturer. And, you need their warranty. The will be their for you. Given that Server 2008 generally runs in the $500 range with five user cals, the server must not be that good. Don't forget about the CALs. I will help you with any question you have with one caveat: Buy SBS 2008 or 2011. I prefer 2008. But, a small business server will give you much more than a standard for a small business. Exchange Server, itself, is worth $1,000. Throw in the greatest, most least appreciated program ever, SharePoint, and you are set. It also comes with SQL Server if you get the premium version.
3. What is the difference between RAID1 and RAID5. I love the wording. If you asked "What is the best RAID, you will get, 2 to the 5th power of answers. In a nutshell, RAID1 is a mirror. Two hard drives which in real time contains the exact same data. So, if one dies, you simply pop in another drive and your RAID controller will rebuild the RAID while you work. You must take into consideration that you lose 1/2 of your total hard drive space, but who cares when hard drives cost next to nothing. A RAID5 doesn't use mirror to be redundant. It uses parity. You need at least three drives, but can use unlimited. Here is how RAID5 works: You have, say, five drives. Four drives will contain your data, while the fifth drive will be your parity drive. It contains no data. All of your data is spread across in stripes across the four hard drives. Parity allows the bits 1 and 0 to use a 3rd bit to provide partiy so that in the event that one drive crashes, not only can your computer continue to run, but by adding another drive, it can be rebuilt. Best article I have seen is
http://riceball.com/d/content/raid-5-parity-what-it-and-how-does-it-work but that is way more than you need to know. You can lose one drive in a RAID5 whether it ia three drives or nine drives. Lose two and you are toast.
You can use RAID1 as your entire setup or RAID5 as your entire setup. Personally, I use RAID1 for my OS and RAID5 for data. A very popular RAID that people are using is a RAID10. A RAID10 takes two mirrors and stripes them so you have two RAID1s configured in a RAID5 config. Here you can lose two drives and be fine UNLESS you lose two drives in the same mirror, which is unlikely. A RAID0 is two RAID5s mirrored, not typically used.
A RAID10 gives you increased performance and redundancy.
One thing I have not mentioned is hot spares and hot swappable spares. Hot swappable spares is a hard drive that is the same size or larger which can be swapped with the bad drive while the server is running. Just make SURE you are swapping with the bad drive. Even better is a hot spare, which sits in the computer and instantly replaces the bad drive and begins rebuilding. You will notice the speed slow down, but it will be rebuilt overnight. You can get a global hot spare for the RAID1 and the RAID5 and one for each. It is a good idea to have all drives be the same size and brand as well as the hot spare. You should have another drive ready to replace the hot spare. I know, TMI.
The advantage of a server/client or server/client domain lies in its ability to give you better redundancy, some increase in speed, less chance of need for maintenance, etc. I don't think you will see a measureable increase in speed in your programs.
