Just when you think it's all going to plan . . .

[maniac]

Basically it boils down to the fact that I've recently been told that people higher up in the tree than me would rather I didn't spend £10's of thousends on a new network setup this year for many reasons, which I won't go into now. It's a pain in the backside, because I've spent the last 3 months researching very carefully what we could change to, and have had many consultants and engineers in to advise and provide quotations for different bits and pieces, we were pretty much at the stage where we could place an order. Then I got told that they no longer wanted to make that investement at the moment. The reasons they have given are perfectally valid and I totally understand why, but I did rather arogantly turn round to the business manager and say 'we've got to, the network just isn't going to last another year in it's current state'

 

Thing is I'm now questioning my judgement, and wondering whether I've been spoilt by previous schools I've worked in having very up-to-date technology, and been very single track minded on wanting to change what we have for no other reason than I 'think' we should. Bear with me a minute or two while I describe my current scinario, and see what you think.

 

We are currently an RMCC3 school, we have 330 Wireless laptops and about 140 Desktops for student use, and 65 Staff laptops, and around 15 staff Desktops around the school. As you can see, it's not exactly a small network. We have RM servers running the main system, with a SIMS server, proxyserver etc but they don't matter as they are fairly new, and would have been kept. It's the mainstay servers of the network that are the issue.

 

Server 1 is fairly new, after its predecessor failed last year, It has around 200GB storage on it, and is a well speced server. However, most of the available storage on it is used as it's the most reliable server on the network, we currently have around 12GB free for staff areas and about 50GB free for student areas (years 7, 8 and 9) on that server.

 

Server 2 is very old, at least 5 years, it's still running server 2000 and doesn't do anything except RADIUS authentication for our wireless system. IT has a habbit of falling over about ever 2 or 3 weeks at the moment.

 

Server 3 is the same as Server 2, and hosts out shared drives, has around 30GB free on it. It also hosts some printers. We have had 2 failed drives in the last 6 months on this server, and it hates being restarted, often hanging on the BIOS screen.

 

Server 4 is around 3 years old, and has about 140GB of storage on it. This server is the second most reliable server on the network, and hosts all our printers, and the year 11 and 12 user areas, as well as running our intranet, helpdesk system, Eclipse and a couple of other SQL databases for other programs.

 

Exchange server is the same age as server 2 and 3 and is running exchange 2003. It struggles, but is actually reasonably reliable.

 

Now my question is this, am I just being single minded in insisting we renew this lot with a brand new setup? What would you do, risk it and make the best of it for another 17 months or keep insisting that the investment was made? I've been assured that we would be able to renew it next year if we don't this year, as we would be able to draw on money that's available for a new building project that's in the works at the moment. I fear that may be too late, but I'm wondering if I'm just over reacting and playing it too safe.

 

The network is a little on the slow side, but does need a lot of TLC to keep it running, like regularly re-booting of servers and regular clear outs of stuff from the kids areas to stop all the space being used up. To the end users the system appears reliable, behind the scenes it's a little chaotic somtimes, but downtime has been less than a day since September. I think because we're so good at keeping it working, they don't see the need to change it because to them it appears to work, and appears to be reliable. I think the business manager thinks I'm being two closed minded about the situation, and that we could 'struggle through' which I protested at when he said it, but thinking about it, maybe I am?

 

(Incidently the proposed new network will be non RM, which saves us £10k a year already on RM licensing, support and warranty extensions for our aging servers. The quotes for it were coming in around the £55k mark, with room to shave off some of the 'would be nice' extras we don't necessarily need.)

 

Any advice appriciated, because I've either got to accept what they've said and make the best of it, or give a damn good convincing arguement (which isn't too much technical jargon about server lifespans and the like) to push the project through.

 

Cheers, Mike

 

And if you've read this far, well done!

[FN-GM]

What do you get for 55k?

 

Did you tell them about the savings with RM? what did they say?

[reggiep]

I'd be interested in what you get for 55k too. I'm about to upgrade our network away from CC3 and I think I'll get about 25k max

[kmount]

I think an assessment like what you've presented us with here would be useful for your school as there are plenty of issues resident which could potentially cause downtime. I won't insult your intelligent so I assume you've tried to resolve the problems with the installations and have determined hardware to be at fault?

 

Perhaps a smaller investment this year in order to sort out the major problems; for example we picked up a Dual Quad Core Xeon with 16GB of RAM and 1TB across Raid5 for £2300 recently from Dell which could probably quite comfortably hold several of those roles you have elsewhere currently.

 

I think you've made a good point and like I've recently had to do before leaving have had to provide a statement of the situation and plausible solutions/repercussions if we don't do something.

[SYNACK]

I would put it to them in examples that apply to them, ask them if they have any six year old computers at their home and ask how quick and reliable they are. Then get them to imagine that computer handeling 100 users at once. Tell them about the failing server and inform them that one of these days it may simply not boot at all and tell them about what that would actually mean for usablility. If one of the servers at that age is showing up these kinds of issues then the others may very well have trouble brewing in them too.

 

I would not think that you are going overboard in asking for this as if they want the network to keep running then something needs to be done. My suggestion would be to do up a secondary intermediary plan with the bare basics required to get stuff stable for the short term and put that to them as stage one to patch the leaking dam.

 

Also work out how much of your time it wastes and add this into your justification as well as the cost savings from going vanilla to put it foward as a business case and show what kind of resources are being wasted currently by the ageing setup.

[jamesb]

You say that they have valid reasons for not wanting the network revamped, so I think you're going to have to abandon the idea of a complete overhaul. What about working out the spend required to keep the network running in at least a usable state for another year? If you go to them with that, costed out, and emphasize that without it the network will be falling into an unusable and unsupportable state they may be able to allocate budget for at least some metaphorical bandages.

[penfold]

You have brought it to the attention of people higher up and have been refused the budget you need. This is unlikely to change, as you have already stated the end user sees a pretty stable reliable network and they dont see it as an urgent problem.

 

And make sure you document everything that needs to be done, you have actually done, and what you tried to do but were rejected. If the budget needs to be spent elsewhere then make sure that you have written documentation that you have requested money to bring the network up to speed but have not received it. That way if the servers do die, then you have covered yourself by stating this was a possibility. Like others have said, see if you can get a smaller budget to tide you over until more money is available. Remember to highlight the fact that servers failing could mean loss of work which could result in students getting lower grades than they could have with a fully functional network.(I found that the only way some people listen is when you identify how this will affect - and lower grades does not reflect well)

[tmcd35]

While I'd agree that atleast two, maybe three, of your server need replacing - by the sounds it, like yesterday - I would question the £55k no wonder they said no! That's almost my entire years budget for everything - Servers, upgrades, pcs, toner, repairs, etc, etc.

 

I'd go back to them with a plan to replace three of the servers with two new state of the art modern servers, maybe a SAN (maybe not) - definately some free virtualisation - ESXi, and bring it in for the £10-£15k mark.

 

With good reasoning and a massively reduced budget they may just be willing to put their hands in their pockets and do what obviously needs to be done.

[srochford]

£55k sounds like an awful lot of money to replace 5 servers; given you say that one of those is only the radius server it's effectively only 4 boxes that need to cost any real money.

 

I'm guessing you're including rather more than 4 servers in the bid you're making so what else is the money going on? It certainly makes sense to replace any server over 3-4 years old (basically, anything which is out of warranty) and I would hope that you can make that point to management.

 

Can you get a quote to just replace the servers? That's likely to be a more reasonable sum of money which the school might be able to afford? Can you phase in the replacements? Not sure when your financial year ends but can you spread it over 2 years? Ideally, if you have a 4 year replacement programme then you want to spread the money so that it's in roughly equal amounts each year - that makes it much easier for the finance people to manage.

[budgester]

How do this all fit in with your existing 5 year rolling replacement plan ?

You do have one don't you ?

 

As you want to replace all of this in one year, it is likely to have an impact in 5/10/15 years.

 

Why not have a rolling plan of one or two servers a year to spread the capital investment and allow the business manager/ SLT to see good planning. I would image they are seeing 55k and he'll probably want to same next year for something else like switches or projectors.

[maniac]

How do this all fit in with your existing 5 year rolling replacement plan ?

You do have one don't you ?

 

As you want to replace all of this in one year, it is likely to have an impact in 5/10/15 years.

 

Why not have a rolling plan of one or two servers a year to spread the capital investment and allow the business manager/ SLT to see good planning. I would image they are seeing 55k and he'll probably want to same next year for something else like switches or projectors.

 

Because we want to loose RM CC3, that's why we're doing it all at once. The network backend here hasn't seen any significant investment for 5 years or so apart from renewing server 1 last year which they were forced to do because the old one died. This was before I started, and I understand it cost about £10k in total just to renew this server once RM had charged all their labour, and resulted in 3 days of downtime, that's the sort of situation I'm trying to avoid.

 

No 5 year plan at the moment, I've only been here 8 months, and pretty much everything needs to be renewed. A majority of my workstations are 5 year old Dell pentium 4 2.0GHz machines, I have 200 Hewlett Packard TC1100 tablet PCs running at 700MHz with 256MB ram that are about as useless as a chocolate fireguard, and basically whole network has lacked investment for so long we're playing catchup majorly. We do have a number of newer machines, but the stuff that's 4-5 years old out numbers the newer equipment by some considerable margin. We also have an expanding role, so the number of machines is going up, meaning all my money is spent just replacing broken hardware, and buying stuff for new areas of the school, but not actually replacing anything than needs to be replaced. Everything is kept until it physically drops to pieces, the kids have no respect for it and the staff hate using it.

 

I want to get the back end sorted out first so that I know that is all good, and then start renewing workstations in batchs in subsequent years. If I know the servers are going to need significant investment again in 5 years time then this can at least be planned for, the problem I've got is there's so much that really needs doing this year due to under investment in the past.

 

Anyhow, I've revised my figure and e-mailed the relevant people. If we can't spend that much, then will they consider spending a smaller amount for us to do the work in house buying lower spec hardware, and sticking with server 2003/exchange 2003 insted of upgrading, but with the expectation that it may not be as long lived as the origenal proposal. We'll see what happens.

 

Cheers for everyone's input.

 

Mike.

[budgester]

Because we want to loose RM CC3, that's why we're doing it all at once. The network backend here hasn't seen any significant investment for 5 years or so apart from renewing server 1 last year which they were forced to do because the old one died. This was before I started, and I understand it cost about £10k in total just to renew this server once RM had charged all their labour, and resulted in 3 days of downtime, that's the sort of situation I'm trying to avoid.

 

Is there a reason for getting rid of RM at the moment ?

 

Have you had RM in for a quote ? Thats your first step.

 

3 days down time, I'm sure ripping and replacing with vanilla is gonna take you a hell of a lot longer than 3 days.

 

No 5 year plan at the moment, I've only been here 8 months, and pretty much everything needs to be renewed. A majority of my workstations are 5 year old Dell pentium 4 2.0GHz machines, I have 200 Hewlett Packard TC1100 tablet PCs running at 700MHz with 256MB ram that are about as useless as a chocolate fireguard, and basically whole network has lacked investment for so long we're playing catchup majorly. We do have a number of newer machines, but the stuff that's 4-5 years old out numbers the newer equipment by some considerable margin. We also have an expanding role, so the number of machines is going up, meaning all my money is spent just replacing broken hardware, and buying stuff for new areas of the school, but not actually replacing anything than needs to be replaced. Everything is kept until it physically drops to pieces, the kids have no respect for it and the staff hate using it.

 

Get your rolling replacement plan sorted, that will actually allow the people with the money to budget correctly, and give you the money you need.

 

I want to get the back end sorted out first so that I know that is all good, and then start renewing workstations in batchs in subsequent years. If I know the servers are going to need significant investment again in 5 years time then this can at least be planned for, the problem I've got is there's so much that really needs doing this year due to under investment in the past.

 

Good plan but have you let Senior Management know that this is what you are planning, and have you documented how you would likely go about it. I notice you say you are only renewing existing workstations, there is no mention of expansion.

 

I.e. once you have X amount of workstations/users you will need X amount of storage.

 

Also will the network infrastructure hold up with all new machines running at 100mb if you have a lot of old kit running at 10.

 

Anyhow, I've revised my figure and e-mailed the relevant people. If we can't spend that much, then will they consider spending a smaller amount for us to do the work in house buying lower spec hardware, and sticking with server 2003/exchange 2003 insted of upgrading, but with the expectation that it may not be as long lived as the origenal proposal. We'll see what happens.

 

Here is a lesson for you right here, you wanted to do some work so you asked for the correct amount of money, in the past people have always been told to go away and do it cheaper (we haven't got the money) so over budget then you might actually get the money you want, or if it's over a certain amount then you are probably required to get 3 quotes, if they are all within 10% of each other then it becomes fairly obvious your not trying to get money for the sake of it.

 

Your basic 5 year plan might look something as simple as below if you have 200 PC's and 4 servers.

 

Year 1

50 PC's * £400 = £20000

1 Server * £5000 = £5000

Licenses £10000 = £10000 (MS, MIS, Software etc)

Consumables = £2000 (Printing)

Maintainence = £2000 (Mice, keyboards, dead HDs)

Total = £39000

 

Year 2

50 PC's * £400 = £20000

1 Server * £5000 = £5000

Licenses £10000 = £10000 (MS, MIS, Software etc)

Consumables = £2000 (Printing)

Maintainence = £2000 (Mice, keyboards, dead HDs)

Total = £39000

 

Year 3

50 PC's * £400 = £20000

1 Server * £5000 = £5000

Licenses £10000 = £10000 (MS, MIS, Software etc)

Consumables = £2000 (Printing)

Maintainence = £2000 (Mice, keyboards, dead HDs)

Total = £39000

 

Year 4

50 PC's * £400 = £20000

1 Server * £5000 = £5000

Licenses £10000 = £10000 (MS, MIS, Software etc)

Consumables = £2000 (Printing)

Maintainence = £2000 (Mice, keyboards, dead HDs)

Total = £39000

 

So thats a bare minimum of £40K a year just to keep you network running. Now if you said £20000 per year for special projects, e.g. whiteboard and projector rollout, network infrastructure upgrade, new ICT suite, laptop trolley replacement, you are looking a figure of £60000 a year.

 

The bursar/school business manager is used budgeting these figures on a yearly basis (this is probably the cost of an Assistant Head) so if they now know they will need to budget £60000 a year for ICT every year and they have the documentation to prove it to governers and the head it is far more likely that you are gonna get your money every year, and it won't seem like you are going cap in hand for £500 every week just to keep a knackered old network just about running.

 

Oh, and see where you are on the FITS framework.

 

You are the network manager, you need to manage it and not firefight.

[synaesthesia]

Probably not worth me sticking my nose in, as you've got most of it worked out nicely already. However, a fair few of my schools have spent something in the area of a couple of hundred quid to turn dog awful 256mb ram'd machines into fairly speedy bits of kit just by going up to 512mb. A tenner per machine (which is expensive as that went through RM) but you could do it for literally a few quid per machine. XP is horrible on 256mb even on a completely clean install with no network, no domain gubbins and a very patient user - especially when you add the likes of shared graphics memory too. It's amazing just how much difference those few pounds makes, leaving the school to spend on other areas that would help more - network upgrades, replacement servers if necessary or replacing the ink in the 100 odd standalone inkjet printers around schools

[Theblacksheep]

Have you had RM in for a quote ? Thats your first step.

 

Get RM to quote for more stuff you dont want?

 

I'm pretty sure they are binning RM for charging 10k to replace each of their own servers with more of the same hardware.

 

5 X 10k = that *extortionate* amount that was turned down. Then theres the yearly costs.... and the extra rubbish new RM server will bring (CC4). Still, on the bright side you maybe able to get it back when CC4 is still not working next year!

Edited April 10, 2009 by Theblacksheep

[kmount]

If you're CC3 currently are you using RM Virus Protect 4? (ie, Symantec). - If so, on the TC1100's it'll be enough by itself to cripple them. At my last place, we had 1 of the original 15 TC1100's still in use merely as a presentation device as it wasn't up to much else now.

[Butuz]

Ok - I am going through exactly the same thing as you - lobbying SMT to spend not insignificant amounts of money upgrading an infrustructure that has been underinvested in for years (yet has remained reliable down to my efforts).

 

The main difference to me and you - is that I have decided to keep CC3 - as it is already paid for, and therefore I am going to spend the next 2-3 years maximising our investment in what is a secure, reliable, and mostly fast CC3 system, rather than ripping it all out and starting from scratch.

 

How are you spending 10k a year on RM support? Here we are spending about £5k a year on RM support contracts for the network, terminal servers, and smarcache. I am not including server warranty costs in this, as whatever network you have you will still need to spend on extending your server warrantys. Personally I think it's £5k a year well spent as I can escalate any severe problems I can't solve myself, and if me and my tech get run over by a bus, the school can call on RM support for help.

 

As for your £55k figure - we need to know exactly what that entails in order to be of any help. If it is pureley for hardware and licencing cost, and you are doing the upgrade yourself then I would suggest you could save significant amounts of money on that - if however it is for consultants / another company to come in an install /set up your new network then £55k could be the price you need to pay.

 

Your right to question your logic and decisions - I question mine when I handed in a plan to spend £20k on infrastructure this year. But at the end of the day my upgrade will provide a faster, more reliable and far more fault tolerant CC3 network, and if this is what the school wants - good on them, of this is not what they want, and they want to spend the 20k on interactive boards - then that is up to them.

 

Justifying your plan as best you can is all you can do - if they choose to ignore your reccomendations then it will come back to bite them in the rear at some stage. Luckily for us - we had a severe server faliure in 2007 which cost alot to put right and entailed over a week of total downtime - so the school are aware of the effect poor infrastructre can have on teching and learning.

 

Butuz

[garethEds]

I agree with most of the points in here - especially the ones about rolling plans. I think replacing all of this in one year is unrealistic. There has to be some sort of priority replacement - maybe this will get you some funds this year.

 

The best way I have of fighting decisions like this is to go with different scenarios to the Bursar/my line manager and outline what has to be done - outline the cost and then present the plan. Plan for Year 1, 2 and 3 - much the same as what has been said on here.

 

Learn how your line managers work - how they think - are they technical? Once you learn how they think then it makes your job a lot easier. I'm lucky, my line manager trusts my decisions and when I explain to him I don't baffle him with science - just the facts and how it will affect the pupils education.

 

Sorry, I'm waffling now,

 

[Edit - oh and yes, I do think it is unreasonable in this climate to expect your school to completely renovate the network - sorry. I understand your reasons for wanting it but with a little bit of spend then maybe things could be improved slightly.]

 

GJE