As someone who worked in email administration, the problem is always how convincing the MS Salesteam are when talking to the C-Suite (I always thought that meant the Crayola people).

Exchange is an OK product but has a number of features which provide increasing lockins to the MS world. It does not quite run to International standards so sometimes mail from platforms that do are not displayed as intended and some outbound mail misbehaves for non-MS recipients. In addition, it must be installed on Windows and uses a very very large number of servers for Enterprise level installations. Also, MS have still not worked out how to so in-place upgrades but since you are running on Windows that becomes expensively irrelevant as you need to replace the OS for the each new version of Exchange.

I work with a lad who had a God/Saviour complex. He's never wrong, he's always right, he knows best. Except he is hardly ever right, and his own prejudices stop him from looking at the problem correctly.

A month or two ago we get an email forwarded to us by our gaffer, and the email is from a Microsoft representative who is our "account manager". They wanted to set up a call to discuss our needs and wants and blah blah fucking blah. It's a sales call. It's a sales pitch. The fucker is just interested in sales, I'm already busy so I immediately say no to this meeting. Except I get pushed in to it by the gaffer along with my colleague because they feel it might be good. I know it won't be, but the lad I work with is eager to point out "This is our account manager at Microsoft so they will know whats best for us". I tell them both that the email reeks of a sales pitch and that is all it will be, because we've never had to deal direct with Microsoft so how could we really have an account manager with them?

I'm overruled.

I'm not 5 minutes in to this meeting and all of what I thought it would be was true. The representative had no real technical background other than what Microsoft has told her to sell. My colleague is going off on mad tangents asking her things she has no idea about because for every question he asks she says "I don't know personally but I will pass this on to the X expert/lead of that department and we can arrange another call to discuss that in detail".

The call, in essence, was trying to get us to shift our licenses to E5. Which is prohibitively expensive for what it provides. There was nothing of value to the call which didn't surprise me at all. When the call finished I went on a call with my colleague and he was so disappointed. He genuinely thought it would be useful and could now see - after the 45 minutes call - exactly what I told him it would be.

I can see though how bigger/larger corporations would fall in to this trap because the representative of the company either isn't cynical enough, knowledgeable enough, or has a true grasp of their own limitations to know they don't have the full picture, they aren't going to be in a position to question everything the seller is saying. They take them on face value, agree to everything, then cannot understand how the picture perfect postcard world Micro$oft offered hasn't materialised.

For public sector there is no real choice except to use E5, or at least E3 + E5 security and compliance etc, which funnily enough costs more than E5.

What is surprising to me is that they don't do an E7 yet to include E5 + Copilot + x + y + z for a few $ a month less than the sum of the components. A lot of orgs would fall over themselves to sign up to that.

At least they are staying with a minimum 12 month contract period, for now.

Not quite true for Exchange now upgrade wise.

You can do an in-place upgrade of Exchange 2019 to Exchange SE whether the OS is 2019, 2022 or 2025.

What you still can't do is an in-place upgrade of the OS underneath Exchange.

So when the OS goes out of support you are still going to have to provision new servers with a new DAG in same org, install Exchange and then move all the mailboxes to databases on the new DAG.

Not a massive issue if VM's, more of a problem if running on tin and org have insisted on using an older version of the OS.

But to get a sufficient number of the best minds, who can envisage and implement a wholesale review, and then actually implement the agreed actions, you need to pay a reasonable amount to attract them from the private sector.

Unfortunately, you also need political leadership that lasts more than [1]eight months to continue to support the objectives and, more importantly, the budget throughout the lifetime of the project. And to have a project that conveniently fits within the remit of a single department.

The government machinery simply doesn't exist to deal with these long-term strategic issues that cross departmental boundaries. As departmental budgets become increasingly squeezed you might imagine it would be an increasing priority to control external expenditure, but much of it will be committed to long-term contracts so the cuts will be made where they can be made quickly.

[1] https://www.theguardian.com/politics/2024/mar/17/cabinet-ministers-last-average-of-eight-months

As a freelancer, I work across Private and Public sector. There isn't enough money in the world to make me want to work in the Public sector.

This isn’t just bad procurement, it is part of the wider rot where government policy is written to serve big multinationals at the expense of the country’s own capabilities. Successive governments have handed over control of critical infrastructure to a foreign, tax-dodging giant, while designing measures like IR35 to squeeze small British businesses out of the market and remove any real competition. They preach “support British business” while shovelling billions to corporations that pay little back and hold the UK in a chokehold. Calling it negligence is generous; if it isn’t deliberate, it might as well be.

When the government commissions custom software the first line of the contract should be "All source code written will copyright UK Government and supplied before payment." followed by something like "Only GPL compatible third party libraries may be included".

That means it will all have to be written from scratch with no code reused from previous work.

For my work, if asked, I put in something about reusing parts of the work.

" It's a truism that moving to open source doesn't save much money, as proprietary licensing spend is replaced by training, support, and local development costs, but reliable data is hard to come by in a field dominated by financial and political interests."

Is it really a truism? What it most likely means is that staff are expected to know Windows and Office without having to be trained and the cost comes afterwards in things like confidential information being CCed, executable files with a .pdf suffix being clicked, Excel being mistaken for a database application and all the rest of it.

When the ensuing disasters are investigated doe "we don't see a need to train staff because we assumed experience of Microsoft products" ever get considered? Perhaps the greatest gain from Mexit wouldn't be the money saved on licences or even the digital sovereignty; it would be proper training.

This truism is brought to you buy the marketing department of the Sirius Cybernetics Corporation…

Open source or not is less relevant than compatibility and sovereignty. Sure, training needs to be considered, but I'd argue that Microsoft's portfolio places an increasing burden on the technical staff and the proposed migrations to Microsoft managed services poses an unacceptable risk to business sovereignty.

<< The cloud and AI, as currently configured, imply and require huge data flows into American owned data centres. Those data centres, even those within UK jurisdiction, cannot prevent US government access. Therefore, the UK cannot consider them suitable for state data.......The US is not trustworthy >>

T H I S says nearly all. And as for leaving AI to the Chinese etc; FFS.

Unfortunately, the history of achievements within UK Government Procurement services, preclude any success in anything at any time in the future, if ever.

I completely agree that globally the majority are stuck with Microsoft however one has to ask what the genuine viable alternatives are and would that be any better,

Just shouting "FOSS" does not solve many of the problems as what works for a small organisation with skilled developers and support does not work when scaled up.

As I see it the only commercial alternative to M365 currently is Google Workspace. I don't see that as any improvement at all and to be honest is a backwards step.

Taking a "Role your own approach":

What is required to setup an email and collaboration system that is used across the entire public sector?

How much would that cost?

Where do you run it - one assumes there is unlikely to be DC capacity now as so much has been shifted to the cloud?

If you do run it in a cloud what are your option?

Microsoft Azure

Amazon AWS

Google Cloud

Oracle

You could go to a smaller provider we are seeing casualties there now as they cannot compete with the convenience the bit boys.

Now we have the office suite, there are alternative and for 95% (maybe more) these will be fine but as soon as you start making exceptions the system breaks down. Moving onto the support. large organisations, public sector or private sector need something where there are contracts, SLAs, guaranteed updates and longevity. The fiasco with CentOS is a perfect example of why just saying "FOSS" or "use Linux" as the generic response raises so many red flags.

Simply saying "replace Windows with Linux & Office with Libre Office" is the easy bit. The reality is very different.

What is unclear in the £9Bn deal is how many users that covers. Education is mentioned and this is absolutely huge as high school upwards all have email. The local school my kids went to switched to Google, ChromeBooks and Google Workspace. That is not a solution, you are just replacing one monstrosity with another.