Planning Out A Network

Ok, I've just had a relatively large project thrown at me. I have been charged with coming up with a proposal and price analysis for a network that will be used by a Municipal Government. Unfortunately they have not really given me much information to work with such as exactly what kind of data they want to store and how they want to do the entry. I am not too concerned about that though. They did not tell me the size of their organization either (I guess they think that anyone's assumption is correct...). It is a small city (about 5,000 - 7,000 residents), and so I am coming up with proposals for around 10 to 20 desktops. I am also including in the proposal an Exchange 2003 Server, Web Server, Storage Server (NAS), and 2 Active Directory Servers. Is this overkill?

For the Exchange 2003 Server I am not sure whether or not it would be best to have a front-end and a back-end, of if that would be unnecissary. If I did have a front-end and a back-end I would put the front-end on the same server as the web server. Here is the hardware I'm looking at for these systems right now:

Active Directory Servers (PowerEdge 2850) - Dual Xeon 3.6Ghz, 1GB DDR2 RAM Single Ranked (2x512MB), 3x73GB 15k RPM SCSI HD w/ RAID 5 (Embedded PERC4ei),  Intel PRO 1000MT Copper Gigabit NIC, Redundant Power Supply, 24x IDE CD-ROM, 1.44MB Floppy Drive, PowerVault 100T DAT72 36/72GB Internal Tape Backup Unit w/Onboard SCSI, Veritas Server Backup Software, Tape Media Cartridge (5 Pack), Windows Server 2003 Standard Edition

Storage Server (PowerVault 770N) - Dual Xeon 3.2Ghz w/ 2mb Cache & 533Mhz FSB, 1GB DDR (2x512MB), 1x6 Hot Pluggable SCSI Hard Drive Backplane, 2x36GB 10k RPM SCSI HD, 6x146GB 10k RPM SCI HD, PERC4-DI w/ 128mb Battery Backed Cache, 2 Internal Channels w/ Embedded RAID, Intel PRO 1000MT Dual Port Copper Gigabit NIC, Redundant Power Supply, Windows Storage Server 2003

Web/Front-End E-mail Server (PowerEdge 1850) - Dual Xeon 3.6Ghz, 2GB DDR2 RAM Single Ranked (2x1GB), 2x73GB 15k RPM SCSI HD w/ RAID 1 (Embedded PERC4ei), Redundant Power Supply,  2x Intel PRO/1000MT Copper NIC, 24x IDE CD-ROM, 1.44MB Floppy Drive, Windows Server 2003 Standard Edition

Back-End E-Mail Server (PowerEdge 2850) - Dual Xeon 3.6Ghz, 4GB DDR2 RAM Single Ranked (4x1GB), 3x300GB 10k RPM SCSI HD w/ RAID 5 (Embedded PERC4ei),  2x Intel PRO 1000MT Copper Gigabit NIC, Redundant Power Supply, 24x IDE CD-ROM, 1.44MB Floppy Drive, PowerVault 100T DAT72 36/72GB Internal Tape Backup Unit w/Onboard SCSI, Veritas Server Backup Software, Tape Media Cartridge (5 Pack), Windows Server 2003 Standard Edition

Is this overkill? I've never really done something this big before. I am kind of unsure if I really need a second domain controller and a seperate storage server. Maybe The second Active Directory Server should also be the file storage server? Or would it be better to have it seperate like this? I planning on them having a LOT of files.

Thanks for your suggestions.
Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

I've been guilty of "over engineering" myself. This does look a little like overkill for 20-30 users. Municipal gov't usually wants everything on a shoestring budget. You might even get away with SBS 2003 and maybe one file server with backup capabilities.

I see no reason at all for a front end/back end mail strategy. You might even consider recommending outsourcing things like spam filter and inline Anti Virus and DR. Take a look at
Use them as your 'front end' and DR
5000 - 7000 residents, serviced by 20 desktops?  That should be your first question.

How about the smaller end of the scale, 1% municipal employees; that would be 50 desktops.

If you don't think that's accurate, think about most governments, the fed has 3 million employees for 300 million people, about 1%.

Take an average number, 6000 residents.  At ten bucks each to get them started, you have $60,000.00  The ongoing budget is up to the residents and their municipal government.

And if it were on a townhall or voters meeting, you could probably get $100.00 each, or $600,000.00

You have to figure your budget first, then build the network.

What you have may seem like overkill, at least at first, but I don't think it will be as time goes by.  As a matter of fact, I think it will be inadequate.

All of your town residents are going to want access to "their" government computers, so now you've got 6000 users.  Administered by however many servers and workstations for municipal employees you have that provide services.  You may not be thinking of the disabled, the elderly, and children, who can't just hop on down to City Hall at the drop of a hat, who will want an internet connection to their government.

You can run it on a shoestring budget, but what you will get is a shoestring network.

Microsoft is very expensive.  While I would say yeah, have a Microsoft Server or two, I would add have also a couple of Linux Servers to back them up.  That is, don't put all of the town's eggs in one expensive Easter Basket.

You have time to learn them both.  And having both will show how competent and smart you are.  Having one or the other, will not.

If you "outsource" any government services, especially without a written order from the Mayor or somebody, guess who gets held liable if government information "leaks" owing to the outsourcing?  You.

This municipal government should have a "Scope Of Work" on that contract; why didn't you get a copy?

A low bid is worse than a high bid.

I've worked for more than one branch of government, in contracts.  There is something called a "magic number."  It's the hourly rate that governments use to calculate the mancosts.  That is, the magic number called the "manhour" rate.  They mutliple this by the total number of hours projected to complete the project.  Whosever contract comes closest, without going under the "magic number" wins the contract.  Like pitching nickels against a wall.

We have a high school nearby that holds 5,000 students, in a major city obviously, how many computers do you think they have?  It sure isn't 20.

A small town usually has more wealth than a large city, and the residents usually spend more, both in taxes and in general.  They expect more too.  With a big project, which is what a municipal network is, you start with the big log and whittle it down, as opposed to gluing a bunch of sticks together to make it work later, which it probably won't.

Ask people in the town what they expect out of this new fangled network in their town, what do "they" want from it.  It's not just the Mayor and city employees that make the decision because the Mayor and the city employees are not paying for it, nor do they own it.  The town residents do.  Survey them first.  A little PR never hurt any contract.  And it will help your career more than sitting down and trying to figure out all the technical bells and whistles everyone is going to throw at you.  If it's your proposal, and your decision, then take the executive command console and start making some executive decisions first.  Which means, don't let some wiseguy take the credit for all of your work.  Get your name known.

You have listed all this technical stuff, which neither the decisionmakers nor the townsfolk haven't a clue about, but you lack some really definitive information on what it is your proposing: taking the town into the 21st Century.  You could talk them blue in the face before they'd understand what a raid controller or a three tiered server backup system is, and you'd be no further than when you started.  Interviewing them is how you get to learn to speak their language.

As an engineer, I could field a lot of questions:

Is the network going to fiber optic to the phone company?
How many connections [employees and residents] can it handle at the same time?
Will there be a lag?
Will it break down?
What am I, as a taxpaying resident of this town, going to get from it?
How much will it cost [answer on a per resident basis first, this doesn't scare people as much as a total]

For these questions, and more, there is also a technical answer that points to the equipment you will need.  Did you tally in the physical construction?  The "wiring?"  How about subcontractors?  Is it in one building?  [that, by the way, is usually not a good idea for government] and more.

What, exactly, is it going to be used for?

Answer these questions, and they will answer the technical questions.

Oh, and I'd suggest you starting writing down everything you're doing on this project and keep a journal, not a computer one, buy about a dozen blank composition books, and at least one ledger.

Get a pocket cassette recorder from Radio Shack to make notes to yourself.

This is not the nickel and dime project your are making it out to be.

Let me know if you need to know how to do something.

Grime121Author Commented:
I understand what you are saying, and I've heard similar things from other people. It seems that some people think this thing should be a relatively small project, while others think that it should be large. I'm kind of stuck trying to decide between the two. I guess it is better to go with the large scale and cut things as needed though.

For clarification let me say some things. I am the only one coming up with a proposal. No one else is even aware of this project right now (including some of the council members). After I come up with some ideas we were going to bring it up to the rest of the council (After the elections as well...). It's not really a typical bid because the government did not come up with the idea in the first place. A guy that works closely with some of the city council members came up with the idea, and so far I am one of the few people he has told about it. I'm still hesitant that this thing is actually going to come together. I am planning like it is though because if I plan like it is than there is a much stronger possibility that it will.

With all of that said, I really only have one question about what you talked about. You discussed the town residents using the network and computer systems, and I'm not quite sure what you meant by that. You said, "All of your town residents are going to want access to "their" government computers...". What exactly do you mean? I don't think any of us were planning on it being a publically available network. The computers will be for office use only I believe. I am interested in understanding what you mean by this though because anything is possible at this point. I have had some ideas about blossoming this network in the future if we get past this stage and perhaps offering a townwide wireless network for either no cost or a very low cost. That is a different subject though. I am open to ideas though so please let me know what you meant by that.
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.

Let me put it more simply:

"I pay your salary, for the city car you drive, into your pension fund, and for the city's computers, Internet, and network, and without me, you don't have a job.  Now, what access am I going to have to this fancy new network you're asking me to buy you?"  quote from a voting city resident.

I don't know if you understand that, but I really think you should.  How any government official thinks that everything should be done in secret, of, for, and by the City Council and City Employees, is a quandery that dates back to the wording of the Constitution, and the references to all government records to be Public Records.  The biggest mistake an elected representative or politician makes is "forgetting."

Every town and city, no matter how small or large, wants to be recognized and be able to say "Look at us!" and "Isn't our town great!"  But more importantly, they want to feel their elected representatives had "their" best interests at heart and include them as part of the family.

I do live in a city of millions, and even here these millions of people all have that feeling of being part of the family, part of their government, and the first thing they did ask for was that all public services be accessible via the Internet, in authorizing the cost.  As a major city, we have to put such things on the ballot itself.  By the way, it virtually guaranteed all re-elections.

Funny, our current IT Head Mistress [the City calls her their network gury; I sure hope she is at $190,000.00] is making the newspapers everyday about a citywide wireless plan, getting everybody onto the city's own little Internet.  Front Page stuff.  She says it's going to be absolutely free for everyone.  So, this is a big sell, big enough that the newspapers are devoting more than a few Front Pages to the issue.

Okay, let's take the average city services, water usually, town bills always, meeting times, social services, maybe power and light, maybe sanitation, services for children, families, the elderly, businesses, residentials and ordinance, legal, court, police and fire departments, schools.  Normally everyone wants access to these from their homes today.  They no longer want to have to make phone calls or a trip to city hall to ask a question which can probably be answered much more quickly on an Internet site sponsored by their city.  It makes sense in a time when there's all this necessity for time management, even in the home.  For example, does little Billy have a homework assignment that's due tomorrow morning, is there a luncheon honoring senior citizens that grandma is going to attend, is the city considering rerouting traffic for construction, nearly anything and everything you can imagine.  People want a central site where they can get this kind of information instantaneously, an Internet site for their city services.  They also want an interactive site, one they can send their opinions to officials on and hope to get the respect of an answer.

And it doesn't hurt the productivity of the city employees.  In fact, most city employees themselves will state that they can do more work because they don't have to go running here and there to find out information, or put everyone on hold when 20 calls come in to their office from town residents, and they can integrate their job functions quite rapidly with the public side of the network.  And there is a public side and a city employees and officials only side.  Apart from the Public Records aspect, the day to day job functions of city employees and representatives can't be always open to public access, we all realise that.  In addition, some records are confidential, at least until published [made public], such as city, fire, police salary negotiations, contract bids, and the personal records of the town's residents [which fall more under Right to Privacy and Fourth Amendment than administrative government records, such as building permits, etc..]

Speaking of building permits, business taxes, property taxes, and so forth, people now expect to be able to go online for such assessments.  They still want all the bills in the mail, as well as notifications, but they also want to be able to instantly check those records, or find out their absolute deadlines, and so on.

I think 5,000 residents is not all that small.  When I think of a small town, perhaps that's  the borderline, but I tend to think of a small town more like one in Texas with 324 population.  It's just personal point of view I suppose.

These people know what other towns and cities are getting, and if you ask them, I'll bet they want in on the same thing.  They want to at least be up to the technological level and sophistication of the other towns across America.  There's also some self pride and self respect involved here.

I have yet to see any city or town population turn down a citywide computer network that they are part of.

It is a large project though, it's a complete change of government, and the relationship between voter and electorate, taxpayer and city employee, Mayor and constituents,etc.. For most towns, it's a jump from the 20th century into the 21st century.  Also, in this case, a jump from one millenium to another.  It's a feeling of starting with success, not ending up there after a long hard struggle.

I could probably write a whole treatise and proposal on this, as I have in the past, but the real point is a time comes when a society has to move forward or fall seriously behind the times.  Not to discount that there are a lot of towns that want to stay where they are, but usually they are wide open ranges, or property rich counties with small populations that think of the city and county as their personal backyards.  The average town in America though is neither of those, maybe close, but not quite the sprawling estates of Senators and Congressmen across the Potomac in what can only be described as the Land of Oz.

Having been elected, a time or two, and having worked in all three branches of government, local, state, and federal, I've come across the public's pulse over many years.  You get to know not only what these people are thinking, but how they think.  And they think pretty much the same way you or I think.  What's good for them, is good for their family, and their family includes their town, from the small town to the big cities, they're not really all that different at all.  You might be surprised to find out that everyone in New York City really thinks of it as nothing more that a "small town" where everybody knows everybody else.  How they can come to that conclusion, I don't know, but I felt the same way while living there.

There are numerous small towns throughout most Eastern States that I have lived in, and felt the same feeling of everybody knowing everybody else.  Strange, but true.  And I can only attribute it to this town feeling of being a family and all its residents family members.  As family members, none of them wants to be left out, left behind, or not part of the process of family decisions, especially those aimed at improving the quality of life within that family.

I have personally been through the transitions from early Burroughs, IBM, Honeywell, Sperry, Rand, Univac, and other systems, to the modern PC and workstation environment for Government Systems.  As well as Corporate and Small Business Systems.  I have worked in both sides of the issue, in government as well as in the private sector selling and supporting systems to all levels of clientele.  From the small home businesses of the Internet of today, to the systems used and employed by the White House and nearly all other federal agencies and the military, Fortune 10 Corporations, and banks, both domestic and foreign, to the level of World Bank.

I was also a design engineer of Very Large Scale Mainframe Computer Systems.  There is one principle that I did learn throughout all of this experience: never put all of your eggs in one basket.  Both from a technical point of view, and from a personal point of view.  Get the advice of the end users first, as well as their opinions and what they think a computer or network is and what it will do for them, and then build the network.

I have literally seen protests and work slowdowns when a computer/network did not do what the end users wanted it to do.  Even a few strikes.  Basically, if it doesn't work the way they thought it was going to work, they won't use it.  They would do everything they could to avoid it, including reverting to printing presses, mimeographs, and other earlier records administration tools.  GM assembly lines would often ignore their terminals and do it manually when the system didn't perform to their specifications.  A reflection on the relationship between man and machine.

To solve this, most network engineers would present a hybrid system from various vendors.  This also helped prop up any failure in a specific vendor's product line.  Although the early examples were mostly Sperry vs IBM vs Burroughs, et al, it quickly became apparent that it was not really the hardware that was the issue, but the various Operating Systems, each originally being proprietary and refusing to work with one another.

With the Internet and Unix, that changed.  Government had mandated that henceforth all Operating Systems would work together or they would not be the speculative venture of government investment, the first big computer customer.

To move forward, to the present day, it is PC's, MAC's, and other software specific machines that now face this same dilemna, and have been facing it for some 25 years now.  While there are various Operating Systems such as Microsoft, Apple, IBM, and others for end users, there is also another choice for providers, similar to the original Unix environment that the U.S. federal government had mandated when inventing the Internet.  Today, that Server Environment is a hybrid of Unix [Linux] and Windows [Microsoft], with Apple, IBM, and others taking up a very small percentage of the marginal market.

What I have advised, from the small businessmen to the large agencies of government, and all of the medium sized businesses and governments in between, is to never go with one Operating System on multiple servers.  Always have the best of both worlds and a backup strategy that provides redundancy and security while providing for 100% Up Time even in the face of catastrophe.  If one Operating System should fail, the other should remain relatively immune, even to such things as viruses, physical attack, and plain old human or computer error.

That being said, it's not that hard to find a multitalented team that can handle both ambidexterously.  A truly qualified IT team and/or manager, should know both of these systems quite well.  And there are plenty of resources for those who will have to learn and learn quickly.

For all Windows [Microsoft] Systems, I generally recommend something on the order of a Resource Kit, such as NT Server Resource Kit, and Back Office Resource Kit, which includes full documentation on "Planning Your Network."

I have one of the references, the Network Resource Guide, on another screen in front of me on another machine.  You can find this quite easily by searching for "Terra Flora" and look for the Microsoft site that presents it, currently at:

This is somewhat larger, I would think, than what you want to do, but if you "size down" the examples, you can achieve your goals.  Understand, that Microsoft is going to suggest only Microsoft because that's how they make their money.  You can find similar planning techniques for Linux, and/or a combination of both.  The Microsoft guide is best employed by considering the dual, or hybrid server based combination of proprietary [Microsoft] and free [Linux]    software.  Doing this from the beginning will allay any concerns about "prohibitive costs" in any implementation of future Server Software vendor migrations [changes from one Server OS to another].

All of your eggs are not in one basket.

The resellers you contact are going to present a lot of arguments for their product, usually Microsoft, but all of these arguments have only one basis in reality, the money they are going to make off the city and its residents.  It has, indeed, made many of them rich.  On the other hand, it has left many cities with cost overruns and budgetary crises after having been locked into a software monopoly that dates back to IBM, then Apple,then Novell, and presently Microsoft.

I am not anti-Microsoft, I am anti-monopoly.

Apple, at one time, had over 80% of all schools and universities business, and if you wanted to attend certain colleges, you had to buy one of their machines.  Not a very fair practice in educating our children by holding the admissions policy and fees of tuition that we parents pay hostage to a surcharge of $2,000.00 [at the time].  When Microsoft came out, the payers - the parents and students revolted and sent Apple into its spiral downward, which nearly made their machine and their software obsolete.  The same has happened to IBM and Novell, it seems to be the ultimate price of such practices.  There is no universal law that says it cannot happen to Microsoft, thus obsoleting most of their current systems.

Microsoft does cost.  I would bet that your network will be assessed at about $100,000.00 in the first year for a full compliment of the the Microsoft suite of programs necessary to implement it.  And the licenses are an ongoing fees revenue stream.  Not that what you get is bad, it is workable and has proven itself, but it is a continuing cost item, it doesn't go away after the first year.

The biggest part of that cost is in the Server Software, including the Operating System and the Per Server or Per Client Licensing structure whose cost increases as the number of users increases.  20 clients at $150.00 is $3,000.00, but that is just for the workstation Operating Systems, most probably XP Pro.  That does not include Office, MSSQL, Access, Exchange, SMS, Lotus, Excel, and a host of other programs that you're going to find out are just as necessary as the Operating System.  On top of that is the fee for help from Microsoft.  Not even an answer to a question as simple as how do I install Microsoft is free.  All technical help and software help costs and this is in a per incident framework.  Workstations, on the other hand, are paid in full up front.  So that no one can charge the users are home for Internet services, such as web browsing, except perhaps their ISP monthly fee.  And no one can charge on a Per Server or Per Client basis for a Linux Server network, its main advantage over Microsoft Server.  Resulting in a more competitive bid from Microsoft vendors and resellers.

Linux has its own suite of programs that not only equate with Microsoft's [this is at the Server level, not at the individual workstations], but most Linux implementations can run many Microsoft programs and applications as free software and free software advocates and developers, such as have "ported" to Windows Servers and workstations.

The point I'm trying to make here is one of cost analysis and negotiable bargaining power; just don't give in to the argument that either Microsoft or Linux is either the better or the more cost effective solution.  Either one alone is not.  The two together, however, are.  Which makes the sum greater than the whole of its parts and a much wiser planning strategy than any single implementation.  And don't let anyone tell you they don't work well together, the fact is that they do, they work better together than either one does alone.  And that is more cost effective and less expensive than choosing one or the other, both in the short term and in the long term.

Which brings us right back around to the taxpayers and residents who expect a fully burp-proof system whose costs can be controlled by them, and not by any vendor or network guru.  They don't need a "guru," they need an advisor and a guide, one who has objectivity and not just proprietary interests at heart.  They already have elected advisors and guides whom they expect to make the best decisions in their best interests, and not in the interests of a select few.

All in all, people are not naiive and they are not stupid.  They do tend to be somewhat quiet about what they know, waiting to see which way the wind blows.  And they nearly always wait for at least a spark before they fan the flames.

By the way, in the Microsoft example, Terra Flora, they may mention things like a "Sun Server" and others, these are Linux systems, in effect.  Even Microsoft recognizes the value of hybrid systems.

Check out the Terra Flora example, look over some figures, perhaps at other cities, and then ask whatever else you'd like to.

I can only add, "Since when has government ever been a small project?"
Councilman, did you get anywhere with this yet?

Do you need some nuts and bolts consulting to get you all the right numbers and a convincing argument pro-networking for the city?

Grime121Author Commented:
Right now we are just sort of testing the waters until after the elections in early April. I'm not one to turn down the help if you are offering it however.

You mentioned that you have seen some of these undertakings bomb, and be very poorly received by the public. What sorts of problems did they run into? What were their mistakes?

My next step is to get some time scheduled in the main Municipal building so that I can look around and see what exactly it is that we are working with. I'm also going to talk with some of the employees and ask them questions such as what their expectations are, what their typical activities are, and what would help them out. Hell, I still haven't gotten a straight answer on how many employees we are talking about. Some people say it's around 5 and some people say it's around 20 or 30. I'm really anxious to get this thing off of the ground, but without these specifics it is making it a little difficult.
The number one mistake made was in hiring the a myopic IT Administrator or funding a myopic contractor.

Which is exactly why I will keep referring to using both Linux and Windows.  You will have two teams that have to both compete and work together.  Without diversity, as Caesar said, there is no victory, diviso est vincere, in division is victory.

In division is also control, as in the council and residents not being duped by either contractor or sly administrator.  The one-eyed or myopic person builds his own fortifications, even against his benefactor.  His advice is useless as it applies only to his personal goals, and not the good of the public interest, and certainly not toward the good of the public.  In flat out political terms, I am calling the "one vendor/one system" proponent a liar.  Not unheard of even in the halls of Congress, and certainly not unheard of in the periphery of the court of the city council.

Because the myopean sees himself as the chief of staff, rather than his employers, be they contracted or not, he sees the people, the residents, as his new little serfdom.  He does not see them as his boss.  This does not sit well with any publicly funded project.  There is first a responsibility to the people, not to the council, nor to the government employees, but to the employer of all of them, the town residents.

The "do it my way or else" philosophy of a one way IT manager creates enmity between the project and its benefactors.  He hires only advocates of his one way system.  He uses them as his rank and file in a private little army, if you will, aimed more at his own struggle for power than at serving the public.  There is no doubt of this if you read a few histories on those who have preceded current one way theosophists; Watson at IBM, his "little colonel" at the Pentagon, his army of employees with secret "hail Watson" meetings in secret basements.  Novell and Apple have similar stories.

The problems arise when such power figures begin dictating to government.  Believe it or not, they have all tried to influence the hiring, firing, and complete staffing of  such things as the federal government.

Not to go too much further into history, which I can assure you is bonafide and back up by facts and documents in many governments, you protect yourself on both flanks.  One Linux, one Microsoft.  That way, neither one can gain leverage over their boss.  They must compete, and in so competing, they must please the public and you, and they must work together.  This is how you balance the scales in your favor.  That competition will cause a catering to all, the Mayor, the Council, the employees, and the town residents, all of whom can see the results.

Never ask a politician how many employees it will take.  They will say 5 when they're thinking 30, or even 50.  Actually, it's your job to tell them how many they need.  I gave the 1% example as a good reference.  It stands up pretty well under any circumstances.  It's just de facto statistics.

About the bombing by the public, also:  what everyone wants, without realising it, is a staff that can help right down to the last completely uneducated user.  That is, the personal approach.  Which means at least one somebody has to do the Public Relations with both employees new to the system and members of the public who want to use the new system.  Users generally do not know the exact things they want or how to get them until they come to the point where they believe it is available, such as after the new system is built and functioning.

Let's do some nuts n bolts.

It looks like your planning 5 servers:

     01.)  Exchange 2003 Server               this is the same as a mail server
     02.)  Active Directory Server            functions as a Web Server
     03.)  Storage Server                     All servers are storage, do you mean repository?
     04.)  Web/Front-End E-mail Server        Internal Web and Email?
     05.)  Back-End E-Mail Server             a third mail server?

All of your suggested machines, had they Thunder type and K7/K8 dual processor with sufficient server slots [7 per board at least], are between $3,000.00 and $5,000.00 - total $25,000.00
These should all be 64-bit machines with all 64-bit devices.  SCSI has 80 pins and until recently IDE could not even carry 64-bit.  The slots should be a majority of 64-bit buses, with the option of, at most, one or two 32-bit slots [which you shouldn't use after a while anyway].  You will have to be very picky about this when purchasing.  Getting the wrong motherboard, wrong cpu's, wrong devices, will bottleneck the network which expected to last at least 10 years before replacement of main servers.  I don't care what others have to say about a computer's life being 3 years or so; that's a sales pitch, not reality, and certainly not cost effective.

So, initial hardware is around $25,000.00 for the servers.  But servers are quite useless without some workstations.  But first, let's do some Microsoft math.

Each Server software is $4,000.00 for 20 client licenses.  Okay.  Then, networking applications, such as Office, Exchange, SMS, full versions of things such as Microsoft SQL, Lotus, and other specialty applications can quickly add up to about $2,000.00 per server, at least, for another say $20,000.00 plus $8,000.00 giving a total of $28,000.00 after initial hardware.  We're now up to about $53,000.00 plus the support fees for Microsoft Support, add another $2,000.00 making it a rounded off $55,000.00

But remember, Microsoft will most likely deprecate these systems and try to sell you all new ones in about 5 years or less.

On the other hand, the Linux Server software is completely free.  Including all of the applications that it runs.  Linux is completely compatible with all Microsoft Server and Client software.  Although it can't, and shouldn't, run Microsoft programs, it either has very similar applications, or, it can call on a Microsoft server when needed.  Apache Web Server holds 80% of all Web Server markets, owing to its better performance and much more able security.

The two will work together.  Linux can auth any Microsoft Server, machine, and/or user with the same credentials that Microsoft uses.  Linux will enable an adept C++ programmer to compile a completely new and super secure Operating System; Microsoft will not.

Let us say you start with 20 workstations.  Budget about $1,000.00 each.  There is no sense in cutting costs by trimming down the actual workhorses of the system and network.  It is not the servers that do the work, it is the people at the workstations.  Nearly all planners get this wrong!  Machines basically "obey."  Workers actually do the work.  Machines may speed it up, but they can just as well slow it down.  Machines are the servants of people, not the other way around.  This is what all employees and users will expect, "obedience of the machine."

Productivity is a function of the quality of the tool and the happiness of the person using the tool.

No one is happy with a bad tool, therefore, productivity drops off, radically.  Using a pair of pliers to loosen a 14mm bolt eventually strips the bolt making it useless.  The proper tool is a 14mm wrench.  This is the analogy of using cheap workstations with highpowered network servers.

Add $20,000.00 initial employee equipment costs; $75,000.00 now.

You have, for a $100,000.00 budget, about $25,000.00 for construction.  Depending on how big the building is, interconnections, etc., it should fall in with a slight excess.  Do not forget about routers.  I think you will need at least two good routers.  My personal preference for such a setup would be Cisco routers, hubs, modules, and switches.

You can't trim now at different places.  Why not one Microsoft mail server, and one Linux mail server?  Either or both can double not only as file servers, but also as Web Servers because you simply don't have the load with even 5,000 resident users to bog any of them down, and with only 20-50 employee users, there should be no heavy load.

The town network should run its own DNS servers and not depend on an upstream ISP.  This means, you need an admin that knows DNS, this is more likely a Linux person than a Windows person.  Any of two servers can be DNS, even if they have to double as such.  The reason I suggested two physical site locations was also for this reason, even if they're only in different parts of the same building.  Often, municipal power is structured in such a way that one end of a building is on a different grid from the other end of the building.  In any case, that is a point worth studying.  But a government is expected to fully run its own network, including DNS, Web Services, etc..  It must remain independent of private industry controls.  Included in that is the first requirement of all security, Physical Site Security.  These security rules are readily available at the N.S.A. and Pentagon sites recommended for all governments:

and elsewhere.

Sorry, you'll have to search for it as these sites have grown considerably and it's getting harder to find things like C2 Security.

I would say that I think your best bet is either two Linux servers and three Windows servers, or three Linus server and two Windows servers.  The workstations don't give a lot of choice with only Microsoft dominating over 90% of the market.  It is odd that this continues as 80% is the definition of monopoly and is where both IBM and Apple were called onto the carpet.

For the residents, you don't need these client access licenses as they are inherent in the purchase of the residents workstations.  A resident cannot be charged any fee for access to government services, regardless of the medium, by private industry, patent holders, and/or copyright holders of devices and software post purchase of the device and/or software.

Basically that means that a private company, such as Microsoft, cannot require a fee that is essentially a tax for use because it is has no rights to function as a branch of government and access to government must remain free access.   This sounds sticky, but it's not really; a private company cannot dictate access to government, quite simply.

So 5,000 residents access to town servers should not even be brought into any question, by any vendor.

There is no cost for the residents network access except to the residents in the form of their phone bills, ISP charges, etc..  There should never be any fee associated with residents' access to their government.

Aside from the cabling in the building, the construction costs already mentioned, there are spurious and oft times out of pocket expenses.  Unforeseen problems, mostly, for which most businesses and governments budget and extra 10%.

This pretty much covers construction and installation.

Thereafter, perhaps with planning before, there are the ongoing manhour costs of operating the systems.  Operations.

That, of course, is up to the city.  Whatever guidelines you use, you will have to at least add one Administrator, preferably two, and a helper or two.  Maybe this is where they're getting the "5" people from.  It will take about 5 people to run the network.  The extra person should also be able to handle all PR and interaction with both employees and residents.  The admins and technicians don't really have time to do this, but the person that does it should be even more knowledgeable than they.  He or she should be a competent admin of both Linux and Windows in their own right.  From experience they will know what is what and what is needed where and when.  And they will also have the communications skills necessary to political life.

The 5 employees are going to cost what the town can afford.  Spread out $200,000.00 over 5 employees however the city sees fit.

A smaller town with access to interested residents and local people can easily trim that down if necessary.  Using one Linux/Windows Admin can cut out one salary, but it also cuts out probably one emergency shift on call employee.

You'd probably be surprised at just how many teenagers in any town, including yours, have already built both Linux and Windows servers.  Were I doing the hiring, however, it would be a fully degreed individual in charge.  The degree preferably being in the field of Engineering, either Electrical, Electronics, or Computer Engineering.  I would not entrust a city government to someone who is either self-taught or two-week-seminar-certified.  Only because experience has shown that those who do not know the rules often cite ignorance of them after they've broken them and have been caught.

And that sounds like a $50,000.00 a year job to me, maybe at least, but it's pretty good for a small town fairly secure position.

In the end, I come up with these figures:

Construction and Installation:  $100,000.00 - $110,000.00
Post Installation fixed variable costs:  $150,000.00 - $200,000.00 per year.

$150,000.00/5,000 residents is $30.00 per year per resident.  Not bad for a city network!

Higher end, just $10.00 per resident per every $50,000.00 increase in costs in the yearly network budget.

Adjust figures as necessary during initial budgetting; but you've got to start somewhere,right?

All in all, that is a great return on investment for the town.  It's liveable, affordable, and reasonable.  It should be within the means of every town resident.

Saving $10.00 on a form that a resident can print out at home is well worth the investment if 5,000 forms have to be filled out yearly.  It lowers the cost for the government and for the residents.  It also speeds up all processes and processing.  Government employees worrying about their jobs being replaced by a computer is an old fallacy; computers result in more jobs, not less.

I'll be around if you need further assistance in April.

Please remember that the problems you asked about were owing to not having a "divised" system.  The eggs all in one basket is the single cause for most failures.

Grime121Author Commented:
My cost estimates were around $100,000 for the initial setup as well. About the Linux/Windows integration though... Active Directory servers require DNS, so both of those Windows servers would also be dns servers. I really don't think there would be a need for more than 2 DNS servers. On top of that, they are going to be used as the file servers (I've scratched the file storage server idea).

The 2 e-mail servers are going to be running Exchange 2003, and therefore I won't be able to use a Linux server for e-mail redundancy because Exchange 2003 most certainly would not work with Linux. Which leaves only the web server. The web server will be running on the same machine as the front-end e-mail server, and since like you said there is really not going to be a very heavy load, there is no sense in buying another machine to run Samba on solely for this purpose.

I would like to get some Linux servers in the mix, but I'm having a little bit of a problem figuring out what roles they could play. The only other thing I could think of was to get another server to run Linux simply for file storage redundancy. However, that would be a lot of money to spend on a server that would be used just for storage.

I've considered using something like Squirrelmail instead of Exchange 2003, but Squirrelmail just seems so crude to me. Exchange 2003 integrates amazingly w/ everything else on the network (web servers, active directory, microsoft office, etc.), and it is definately feature-rich and easy to use. I am very impressed by it, although the pricetag on it is a little bit steap.

I'm not too clear on how exactly the licensing works when it comes to something like Exchange. if you get a 25 client license, does that mean you can only have 25 mailboxes or that only up tp 25 users can be connected simotaneously? I imagine that it is per mailbox, but I just want to make sure.
Grime121Author Commented:
Oops... I said, "The web server will be running on the same machine as the front-end e-mail server, and since like you said there is really not going to be a very heavy load, there is no sense in buying another machine to run Samba on solely for this purpose." Obviously I meant Apache.
Agreed.  Web and email should be fine together.  Dell, by the way, has a nice Terabyte file storage server; do you know your capacity?

Personally, I don't see the need for two Exchange servers.  I would use Sendmail on the Linux Server for external mail and Exchange on the Windows Server for internal mail, which I do and they work together.  It does get a little tricky with authorization, but it works out in the end.

The Linux Server can also run Samba, believe me, it can take that load and a lot more.

Active Directory does present a heavy load.  It's okay on one server, I would think, but I would avoid dependency on it and the .net servers.

25 Licenses means you can have 25 machines on your Windows network.  It doesn't have much to do with email.  You should be able to assign as many emails as you like.  The advantage of integrating Linux is that you can avoid any and all conflicts about such licensing by using Linux for the purpose of handling any excess and/or increase in load.  When you buy a computer, it comes with a license anyway.  The Windows licensing scheme is more related to their support pricing than anything else.  They would look very foolish indeed trying to charge a government per employee, which starts to sound like a tax on government by private industry.

So, it's really about supporting different sized networks.

Linux can do nearly anything with Windows, including serving files, backing up all three formats of Windows type files, FAT, FAT32, and NTFS.  It will stay up when Windows crashes and provide more than backup.  I use Linux to troubleshoot more Windows problems than using Windows because it's often hard to troubleshoot a machine that won't come up with its full operating system in tact.  Linux can keep an alternate copy of the Windows Operating system, in case of severe crash and recovery is needed fast.

I run two Apache Web Servers on Linux, one External, one internal.  And one or more on Windows using IIS.  Linux catches the external email, as I said, and forwards it first to an interim machine, that is not actually a server, for viruses and the like, if the email is intended for the internal mail server.  Basically filtering it using a simple XP console that is also one of the admin user's desktop.  There are many methods of doing this.  But the servers themselves should not be assigned the job of virus scanning, it's too much of a load.  Internal email, interoffice and such, shouldn't have viruses, but each machine should have an antivirus program running that includes email checking for viruses.

I personally make the Linux Server the DNS Master Zone because it has more control and less vulnerability than Windows DNS.  Sometimes a government has more than one domain name.

"" can be easily assigned to one, getting a second domain name may require an official registration of one such as ""

With that, their should be more than on Public Static IP Address and more than ISP connection to handle them.  In general, their should be one Public Static IP Address per domain name, at least.  So, if you have two Name Servers and four total servers, you should have 4 Public Static IP Addresses and four discrete lines to whatever ISP(s) your considering.  The four lines would normally go out in different directions to different phone substations, if possible, and there is really no sense in them being on the same trunk line at any point.  The whole idea is to keep the network up and you can't do this very well if everything is in one cable and that cable gets broken for whatever reason.

But even with only two Public Static IP Addresses you accomplish this, and have the network always up even when three out of four servers go down.  You can't do it with only one.

If you lock yourself into out of the box proprietary applications too much, the town loses its leverage.

You will have to have someone that knows how to work with other Open Source or your network is only going to work as long as the software company is alive.  And governments are supposed to last longer than that, well, at least somewhat longer.  The point is while it seems easier to go with the plug and play stuff, it isn't, not in the long run.

If Squirrelmail goes out of business in two years, all of your email will be very costly to convert at that time for yet another new program.  A lot more than the price of the program, like the price of a consultant, which can get very expensive indeed.

You could ask any Linux Administrator what role Linux plays in a Windows network and he will tell you that it keeps the Windows network running.

It's the Captain of the team, can throw the ball, catch the ball, block, run, and make touch downs while the Windows fullback is lying in the mud, pretty much describes its job.

The price you don't pay for in software for Linux can then be invested in a very smart human being who knows more than the certified Windows administrator.  Star quarterbacks get paid more than second string only because they're better.  I don't know how else to put it.

I could pick up a book and learn any system in Windows in two weeks.  I could then pay my $100.00 or so and get certified.  But going to a university and graduating costs a lot more and I come out with a lot more knowledge, thousands times more, and it costs a thousand times more!

And you will get what you pay for.

Linux people tend to know more about servers, programming, scripting, and definitely hardware and software.  That's just the way it is.  You have to have someone that has been taught, at a university, how to think, not just how to go to a book or a help program and look for an answer that someone else has already written about.  All of the answers don't exist yet and weaning someone from dependency one preordained answers is extremely difficult, at best.  You need at least one thinker on the team, and my suggestion is that he be the leader.

The per user licensing was a Novell idea; you can see where it got them.

You won't have to worry about that if you use a hybrid network, and there's no better suggestion than that.

You're real worries are going to be in actual construction; that is where the pressure is going to come in from.  Getting the wires in the walls, re-sheetrocking the walls or whatever while employees are working, and getting it done in a timely manner.  You need a really reliable and knowledgeable contractor for this.  Not just some guy that says he does it all the time, or someone who says he knows how to install CAT5 or CAT6, but someone that really does it all the time, like phone company employees who also have their own business installing networks in buildings, etc..  You don't want a residential contractor that builds or refurbishes houses, you want someone that really knows how to do this.

Have you asked the phone company about a fibre optic line [one or more] yet?

You need to know this for the long term future, for the prosperity of the town, and the requirements of contracting the construction.

CAT5 and CAT6 wouldn't seem to be enough to me.  Yes, inside, but not for township and world, or whatever, presence.

Is the phone company installing already or planning to install fibre, and do it underground, within the township, and when?

So, if the township doesn't have a Civil Engineer, you should also call or talk to a Professional Engineer [PE].  Every state has them, even on the state payrolls.  They'll be very glad to help you.

It won't cost you anything to ask them a few questions.

This is just part of city planning.  You may not have a city planning board yet, but you can always get help from others in planning your physical network.

And the physical network is the first network you have to plan.

Which makes you the developer, a very responsible position, beyond sales and vending, more like the producer and man responsible for the whole thing.

And, just like I said, cause for a whole lot of footwork.

While you're at it, you should budget for AutoCad, the real one, for Architects and Engineers, the town will need it, and they'll love it.  About $4,000.00 for a government at:

for a network version.  Most of the good programs, like AutoCad, should do things like check the NEC codes, building codes, specify all tolerances, in general, and some even do the estimate for you.  All the while making the blueprints for you.

Having done all of this, and by hand long before computers could do it, and having worked on computer designs, I truly appreciate what these things can do today.

I still sometimes draw and do things the old way, just to remind myself that it was not a machine that made all of this possible, but human beings, and it is still a work of art in progress.

And more than anything, that is what your network should be, a work of art in progress.  Not "just another network."

Great works of art, and great works in general, are accomplished by many people under the leadership of a Master Craftsman.  It is the involvement of the people that makes a work great and stand as a testament and a tool.  It is never one person and certainly never one company.

So, if you're going to take the helm, take the helm, for better or worse, it is your ship, after all.

Choose your Command wisely, try to look beyond and see the big picture, and get at least a couple of captains with bonafide credentials if and when you get the contract.

Keep asking questions if you like, about costs, whatever, I put a lot of this stuff in the books I write.  I don't do this just for the "points."

And whether it's Apache or Samba or Swat, or a dozen other programs, no need to worry, your system can run anything and everything you need with no problem.  They will all fall into place with the right team structure and a flexible approach.

You still need to do also a Break Even Analysis, just like any business, and find out when the costs will settle down and what they'll be in a year, two years, up to five years; perhaps longer for government.

And think of it this way: the fast food approach is not going to support any community.


Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Network Analysis

From novice to tech pro — start learning today.

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.