Cloud vs Physical Server

[DSAGEY]

Hello,

 

One of our Govorners want us to go Cloud based and get rid of the physical server.

 

Has anyone else gone down this route of not having a physical server? and using Microsoft Azure instead? Are you able to deploy software such as SMART Notebook, TargetTracker etc?

 

Thanks

[mavhc]

Ask them why they want you to do that, and for their cost/benefit calculations (I'm assuming it's just some random person who's heard of this new computers in the cloud thing)

 

You can rent a server, but deploying large applications would be slow, tell them you'll need to get faster internet for starters, and redundant connections

[DSAGEY]

They want to do it to save costs of the support/off-site backup.

[nicholab]

So are we taking a managed service? So you would need to factor in an MDM to mange your machine and moving file storage and MIS. Also the cost with all of these.

 

In the cloud you would need:

Directory/Identity/AD

File storage/Sharepoint

E-mail

MIS

Printing

Software deployment/MDM

 

In business environments the use of MS E5 is quite common as that gives you all you need. It also works well with a one to one operation like you have in a lot of business.

[FN-GM]

 

One of our Govorners want us to go Cloud based and get rid of the physical server.

 

It's not the governors place to get involved with operational issues....

 

 

Are you able to deploy software such as SMART Notebook, TargetTracker etc?

 

You can help these via Intune.

Edited May 9, 2023 by FN-GM

[localzuk]

Moving wholesale to cloud requires significant planning. You would need to identify every function your onsite solution provides, and then ensure the cloud replacement can do all those things. There are potentially also training implications for staff should any of the cloud replacements be in any way different.

 

We are slowly but surely migrating all our onsite services to cloud services - but only when we're sure the cloud solution can provide everything we need at a sensible price.

 

Simple lift and shift is not cost effective at all. And certain services, when you look at cloud provider options, can be significantly more expensive than running it yourself (eg. for us, moving our VOIP phone system to a cloud provider would be extremely expensive vs the price we pay to run 3cx onsite).

 

"Cost savings" are not a given for such a project.

[HPlum78]

I have posted regarding this in the past look em up, but the TLDR is essentially down to flavour of cloud. By that I mean if you are going to do a lift and shift servers to cloud migration IaaS then you are not going to save a penny (if cost is your only driver) if you run web servers then migration of these to PaaS you should realise a saving. Then there is SaaS where you should be able to garner a better saving again.

 

But that is if you are solely basing you cloud journey on cost benefits alone, there is lots more to consider when thinking of cloud migration.

[HPlum78]

https://learn.microsoft.com/en-us/azure/cloud-adoption-framework/ready/landing-zone/

 

The above is useful for planning your cloudy environment(s).

[Michael]

In my experience SLT/Governing bodies are most interested in exercising cloud storage for data as their primary reason for reducing the requirement for an on-site server. Reducing and removing of course are two different processes.

 

Target Tracker as a good example is a legacy application, but saves Assessment data in the cloud - very much hybrid, but still requires a server to deploy the regular updates it requires. Longer term of course changing to a completely cloud based Assessment tool or one built into a modern MIS is of course desirable, but I consider this more a curriculum/joint focused decision and not solely ICT manager.

 

Once these processes/facts are highlighted, then an action plan/strategy can be formed keeping everyone on board.

[Aprice]

As others have said this would require extensive planning, and possibly changing the way some things work / are used onsite.

 

To do a like for like replacement of the server using virtual machines in Azure would be very expensive.

 

You could speak to Advanced IT Services, and get an idea of what they would charge to lift and shift to their datacenter, I suspect it would be considerably cheaper than virtualising in Azure.

[speakercon]

Don't forget once you have gone server-less those pesky server and comms rooms will no longer be required so they can be used for more offices Edited May 9, 2023 by speakercon

[chaplic]

Hi From Suffolk!

 

Did this pre-covid (just) and never looked back. The office 365 offerings Microsoft gives schools is pretty generous and will take care of most of your file storage needs, intune covers PC management and software deployment.

 

I would say you probably won't be using azure as such; hosting your existing server there is possible but makes little sense. We still have SIMs on-prem because there's not a better alternative.

 

I'd say there is a step-up in technical knowledge required and some things are just not as slick but having laptops vpn-less, access to all data and able to be managed is great (also makes DR plans very easy).

[Jcx500]

I wouldn’t just lift and shift tbf, either move to cloud native solutions such as AAD or move your services to a collocation or a similar solution

[Brummytech]

We went intune route so all of our schools just have legacy servers that handle dhcp/dns, kms,Paxton server and backups for sign in systems. All our Apps are just employed via endpoint manager (Just get msi and convert to intunewinapp or MSI and then add groups as needed. I cant see the need for a azure server.

 

We are looking to move dhcp to the router (Which is managed by isp) and move some schools (Smaller ones) to a onsite 1u desktop just for paxton then use mak and redstor on sign in consoles to cover the rest - but obvs a work in progress and servers are paid for and underworked. I think Primary wise its fine the only issue we have is with students saving work on onedrives and then logging before it gets synced you can work round this with either 1-1 devices or a mapper.

 

Governors who know a little/lot can cause quite a few headaches

[mavhc]

The long term plan would be: get faster internet, and 2nd internet backup. Get 1 to 1 devices for all staff and students, replace all legacy software with cloud software, and then pay for it every month forever

[foofighterjim]

Just playing devils advocate, the lift and shift method may have very short term benefits in specific scenarios. For instance, you may be progressing towards serverless but your sever hardware is due a refresh in the immediate future, it may then be a reasonable consideration to run like this for 12 months to allow time to move any remaining services to native cloud, rather than investing in hardware that would be all but redundant after 12 months.

[Marshatron]

I have been asked something similar but this is only for a one and half form entry primary school ~300 pupils and ~40 staff (~120 Windows devices) The on-prem server is pushing 8 years old and I've been asked to explore the cloud costs vs replacing the on-prem server. Staff already make heavy use of SharePoint Document Libraries so the server essentially just provides AD/DNS/DHCP etc. I don't want to host a server in Azure, I would like to manage via Intune, if possible. They use MS Education A1 Licenses but I don't believe this includes Intune? Would we have to upgrade to an A3 license? And this would be for every pupil and member of staff or is it per device? It's early days and I'm still trying to get my head around it all, any advice would be gratefully received. Thanks.

[chaplic]

If you have A1, I’d suggest the minimum you need would be Azure AD Premium P1. You can get that on a ratio of something like 1:15 FTE to pupils and it’s quite decent. Gives you intune and Conditional Access

 

In terms of your own development, get yourself a O365 developers tenant and start doing autopilot!

[djm968]

Hello,

One of our Govorners want us to go Cloud based and get rid of the physical server.

 

Been there several times before with many Schools. All they see is the headline grabbing "cloud" as the way forward and think you can just get rid of the server(s) (along with the outsourced support contract) Unfortunately, it isn't quite that simple. As others have already said, it takes a lot of planning and time along with the associated costs to implement the move to the "cloud". (unless you have the resources available do this in-house). It also often requires a change in mindset around the the way you work and this isn't going to happen overnight!

 

If you are already using 365 or Google, and you haven't already migrated any data, investigate the options around moving data from the on-premises server to these "cloud" platforms as a starting point. You can tell the Governor we are using the "cloud" for xyz and planning a full migration.

 

The "cloud" isn't a silver bullet, you have to invest in the infrastructure to support it and fill the gaps, before you can throw the servers in the bin.

[mavhc]

Main reason for schools is it's mostly free, stop paying for storage and Office, use the web versions from MS or Google, plus they have way better backups than you do to guard against physical failure

[djm968]

It's early days and I'm still trying to get my head around it all, any advice would be gratefully received. Thanks.

 

With regards to licensing, get whoever you currently use for MS licensing, to do the legwork for you, if they won't go to someone who will. You need to do a full audit of all software and services that are dependant on the existing server(s) all functions that the servers are currently operating and find alternatives or "cloud" options. What MIS are you using?

 

Moving to the "cloud" shifts the storage and processing of data from the on-premises server(s) so you also become more dependant on the infrastructure and your Internet connection for access to your data, and instead of server hardware, this is where any investment will be required. As a minimum, you will need to think about redundant internet connections and UPSs for your main data cabs etc, especially if you are going to cloud hosting for VoIP.

Edited May 12, 2023 by djm968

[localzuk]

Just playing devils advocate, the lift and shift method may have very short term benefits in specific scenarios. For instance, you may be progressing towards serverless but your sever hardware is due a refresh in the immediate future, it may then be a reasonable consideration to run like this for 12 months to allow time to move any remaining services to native cloud, rather than investing in hardware that would be all but redundant after 12 months.

One thing to take into account is modern server architectures.

 

For example, we refreshed our server cluster recently. Instead of buying new servers, we bought 3 refurbished servers, each with 192GB RAM, dual 28 core CPUs, multiple 10GbE network cards etc... Then put some Micron Enterprise SSDs in them for the storage, and utilise hyper-converged architecture - so Hyper-V and Storage Spaces Direct on them. Entire cost of the refresh? Less than £10,000. Servers have 3 year warranties, and we're more resilient than our old more traditional servers + SAN setup, as we can theoretically lose 2/3rds of the cluster without losing functionality.

 

This compares to roughly £70,000 for our server refresh the previous time around with brand new kit.

 

So, its possible to get massive cost savings, whilst also not moving to cloud. Lifting our existing workloads into Azure would've cost us more than £10k in fees in the first year alone.

[HPlum78]

I have seen this a number of times before and its around the redundant internet connection risk, we dont operate a redundant internet connection and have been using cloudy based services of all flavours for well a long time (since the Azure portal had 3 things in it lets say!). Is internet connectivity so unreliable these days? or is this related to an individual supplier? is it not better to write up the risks and accept those risks (and yes I understand that those risks are not the IT departments risks to sign off on) than paying for a whole load of redundant links? It would be interesting to here from those that have/ are cloud only environments on this subject.

 

To be clear I am not calling anyone out here, just interested in how this is being assessed and how big an issue this is in this space, also brings me on to the point I was about to make around even MS have stopped saying migrate everything to cloud and Hybrid cloud is going to be the operating model for the foreseeable future in most organisations.

[mavhc]

UK school gets a budget of £7500 per pupil, so £37 per day.

 

Worst case is someone digs through your fiber line, and it takes a week to fix, chance of happening, say, 1 in a million days

Then you have it fails for a day, say 1 in a 1000, based on my personal data

 

So that's 3p/pupil/day, so for a 1000 pupil school that's worth £37/day or £740/month, if you assume without internet nothing at all can happen in the school and it might as well close, definitely worth getting a DSL/cable backup line to keep something running

[djm968]

If you assume without internet nothing at all can happen in the school and it might as well close, definitely worth getting a DSL/cable backup line to keep something running

Absolutely this. It all comes down to managing risk and considering the relatively small investment and costs involved, I would suggest a diversely routed backup DSL line should be included as part of any risk mitigation plan.