Small Business Server

note: planning the network is the biggest and most time consuming task. The time for the installation and setup of the entire network really depends upon how well planned the network is.

I gave 2 to 3 weeks for 1 person working 10 hours per day, installing from scratch [includes running cabling, switching, routing] but does not include building the computers from scratch [assumed prebuilt computers with no O/S installed].

This also assumes you will install the O/S on workstations across the network using either attended or unattended install of W2K O/S for the workstations.

Maxwell Edison, Lordinfidel, TheChas, Anna are all the top TechRebublic answerers and probably can give you top notch advise. I've installed SBS with AD, Exchange, ISA with 30 Workstations, IIS with a web site in three weeks two men. 4 days of that was just planning out the topology of the network and getting mgt approval. 2 days planning the physical layout. 2 days of user needs, 2 days setup of the PDC BDC. 10 hours of deployment, service pack installation. 2 days network test just to verify everything worked, 2 days network load testing. 3 days of exchange server fumbling, 2 days of ISA server access rights setup. Then there was ongoing tweeking because users would want certain features that were not in the original plan etc etc

Maxwell has a valid point in the amount of time I estimated, believing about half the 2 to 3 weeks time I suggested along with my comments on the # of days I suggested in my breakdown. I'll try to elaborate on some of the problems we had that took the most time. SBS Server PDC setup. That was a pain. We installed this on a computer using an AMD XP1800 processor, 768 MB DDR memory, 2 ea Quantum Fireball AS 20 GB hard drive in a Mirrored RAID array, with 2 NIC cards [note SBS Server has Exchange 2000 and ISA 2000 servers which will be installed all on the same computer as 2000 server. In setting all this up, 2 NIC cards, 1 LAN and 1 WAN was the easiest approach.]. We found that with all the software residing on the one server, 768 MB of memory was required[we ran into low memory problems with only half a gig] Also, all that software took up almost 6 GB of space on the hard drives, not including the 200 MB Cache or ISA server on the partition and the 100MB cache or the DNS also in the partition. continued:

6GB doesn't seem like a lot but in installing SBS, the recommended cache locations was on a different drive that what the O/S resides on. Therefore, we partitioned the master and left the mirrored alone. Second most biggest problem was Exchange Server. There is something about DNS setup with your domain name [using SBS software] where a wildcard [*} domain name is made in the foward lookup zone by default along with the domain name you choose. That wild card {*} in the foward lookup zone of DNSmakes Exchange Server not work. [took quite a while to figure that one out]and we figured it out by accident. I looked at the DNS and wondered why we needed a wildcard in our foward lookup zone. We didn't so I just deleted it. All of a sudden Exchange worked. [see Technet article Q309793]. ISA server was the third most problematic. With Active Directory installed, ISA server, by default, denies any access. Configuring all the packet rules and tweeking ISA server in conjunction with Active Directory groups was very very time consuming as we had to test each packet filter assigned to each group with the workstations for each group. Documentation of the network as time consuming since this was an outside contractor job and the inhouse IT person(s) were to just maintain and use Active Directory. [we told the owners that was not a great idea but they were insistant in not having to spend salary on knowledgable people to maintain their system]. The last was post installation support. Post installation support seemed to be the most costly item [not time consuming during installation]. cont:

We guarenteed out work and provided post installation support both via telephone and on site tech support [if required]. we were virtually on the phone with them everyday walking the IT tech though what constitues a problem and what doesn't [exchange server may make the task manager show 100% cpu utiliztion which is NOT a critical error, just intensive processing] and to NOT shut down the server or reboot the server. after about two weeks of that, we decided that remote access administration over at our place was a better idea than phone calls [which the owners reluctently agreed to provide us with a support admin contract]. Worked better that way cuz their admin would just make up an account in AD then email us with whatever rights mgt wanted and we setup their Exchange server and ISA server over at our place. I learned long ago that in estimating something, double what you think it will take cuz there is always something that will come up. If you follow Maxwell's advise on contracttype and pricing and come under budget, they love ya [and you've made $$ cuz you padded it to allow for what if's or how come we didn't think of that].

Further, I said the Server was the most problematic and that stemmed from having to set up the whole thing and finding out that a static IP arraingement wasn't going to work and moving everything over to DHCP. That the owners, [god bless then] wanted a private internet access for their laptops [we always wondered about that] but didn't want to add DSL lines, so we used an 8 port DSL cable router in between the ISP and ISA server on the WAN. Users hated their cubicles and would move to another [for some reason we never understood] and would try to take their computers with them by swapping out what computer was in the old cubicle with their stuff. We told the tech on site that that really defeats everything the owner set up for security. Heshrugged his shoulders. What we found out was people would learn what cubicles [empty ones we ran cable to for expansion ] had internet access and would just move themselves over there. All in all, it was an interesting job.

Poster rated this answer

Having said all that, I would suggest that you break down all the individual tasks/considerations, as DR The Corporate Groups illustrated. Make your own list, and try to cover all your bases – not leaving out ANY necessary tasks – no matter how small it may initially seem. Then assign a time and money value to each task. Add up the values, add on a “margin of error” factor (perhaps 10% - 20%), and you have your estimate.

However, I might suggest one thing. From the client’s perspective, it’s preferable to know the maximum amount the project could possibly cost, but still have some hope that it might cost a little less. From your perspective, you want to make sure you’re not working (too much) for little or nothing, and you get fairly paid for your time. Therefore, a time and materials estimate with a “not to exceed” maximum ceiling might be a win-win scenario. The client knows that he/she is paying you whatever $$$ per hour plus whatever $$$ you spend for materials, but the total will never exceed a pre-determined amount. This will also help deflect any “padding the clock” concerns the client might have. On the other hand, you would be extremely motivated to complete the project on time, or else your dollar per hour rate will begin to decline. (However, it will never decline to a total zero.) This approach might also help you secure the job in the first place, since it provides both flexibility and a “set in stone” bid.

Best of luck to you.

Poster rated this answer