Dynamics 365 for Finance and Operations On-Prem

/ Kurt Hatlevik

Now the On-Prem system requirements are available : https://www.microsoft.com/en-us/download/details.aspx?id=55496

To get it running the minimum recommended requirements are:

Total Number of instances(VM/Machines) : 21

Total number of CPU cores : 104

Total Memory : 408 Gb

This minimum configuration will estimated be able to support 240-1200 users.

To sum it up: Go Cloud . Much smarter.

16 thoughts on “Dynamics 365 for Finance and Operations On-Prem”

  1. Hi Kurt,

    I would like to know how you came up with these numbers?

    Total Number of instances(VM/Machines) : 21
    Total number of CPU cores : 104
    Total Memory : 408 Gb

    I didnt find mentioned these in Microsoft document.

    Thanks

    Like

    Reply

    • Hi Jeff. The numbers are based on the minimum number of machines specified in the document “Dynamics 365 for Finance and Operations, Enterprise edition (on-premises) system requirements (3).pdf”, and then summed up with memory and CPU’s per machine. All is in the Microsoft document.

      Like

      Reply

      • Hi Kurt,
        I am looking D365FO On-Premises testing environment.
        Could you please share me minimum recommended requirements

        Like

  2. Hi Kurt,
    While the discussion in the comments thread is going into another direction of Cloud vs On-prem debate for ERP….let me ask you a very basic question. To you knowledge, this configuration is the “bare minimum” configuration to have or you can go below that for a lesser number of users (i.e. approx. 50)…..I am working on quoting capital expenditure to a client and this is one of the top question because they are a small company (not more than 60 users) and they (for now) don’t want to go on Cloud.

    Like

    Reply

      • Well….may be they end up choosing the cloud 4-5 months down the road, but they are not sold on cloud for now, so I wanted to know the bare minimum requirements to host 3 non-PROD instances ( i.e. TEST / UAT / STAGING) and 1 PROD instance (of course) — they wanted to know the costs up front.

        Like

  3. There are pros and cons of everything. its not just about cloud, its also about continuous code upgrade, about subscription pricing, and as well as how different company’s view opex vs capex decisions, and what they consider to be an outsourceable expense and what they see a strategic.

    Its also how much control companies they feel they have in a world of digital disruption and whether they can take sensible investment decisions
    How many changes to product direction and technical and code architecture have we seen in the last year -on premise/not-on premise and what does it really mean?
    How many isv module providers were prepared for the hard sealing of layers. Just how fast are they expected to upgrade a vertical solution that has many man years of development? Let alone partners keeping up with statutory changes to taxes and payroll across multi geographies every year.

    Some cons of subscription pricing:
    Budgets approvals are difficult when costs e.g. for usage are uncertain, and there annual unknown price increases, which may be for new features not needed.

    There is an economic cycle companies do have lean years- cash flows are impacted by other businesses over which you have limited control – if you cannot make your next subscription payment you lose your system and then quickly go out of business. When times are tough if you own the license then on premise you can decide not to upgrade without getting a price increase, you can decide not to give a pay rise, or not to pay for support etc.

    Scalability, may not matter much to a mid-size company.
    Varying numbers of users affects some companies more than others – in practice it will be like mobile phone contracts it takes a long time to change.

    Regular monthly upgrades offer huge benefits in functionality but how fast can I absorb and test these across multiple deployments with different taxes, payrolls and lines of business? The there is a need to make any code/report/interface/security/ workflow adjustments, then I to train and roll out and…… hello…… is that the next months’ updates I see arriving. Under the current support strategy (whether on cloud or on premise) you need to update at least once year or downgrade to Ax 2012– (and for how much longer is that supported?)

    Hoe many customers that have moved to the cloud to outsource their IT operations expect to maintain an on-premise team to continually check code upgrades and make any changes needed and then promote those changes through LCs each month? That seems to the expectation. Are partners expected to test the new code every month for every customer build? Do customers have any control over how much time that will need? or expect to pay for that monthly service?

    Code changes for a single company are fine. Moving code through LCS for a global deployment which need to synchronise across multiple users/partners is already a major challenge that holds up projects- and code moves from test to production might typically be one a week at best further delay to wait for Microsoft, to do its thing is more challenging than seems to be realised.

    The evidence is that customers generally like the web UI and the CDS., Power Apps, Flow Power BI story but there is also little sign of them flocking to buy or to upgrade. Many have waited for on-premise and now find there are major infrastructure costs, no retail, no Power Bi/Power apps etc

    The significant license price increases, during what is still a global recession, are also a factor compared to huge discounts for on-premise licenses from other providers. Many see cloud as much more expensive. They don’t get the expected labour savings, they still have clients pcs, networks and printers and interfaces to manage.

    Not all partners can change over night. Labour laws are not the same everywhere, nor are available skills. It takes time to retrain. Cashflows are now quite different. Retraining costs are high.

    The new MCSE and MCSD certification requirements make absolutely no sense. (A finance consultant needs to first qualify as an SQL DBA (why for azure?) and as a programmer which are quite different skills and we have not got to the finance part yet?)

    Partners whose focus is business consultancy with expertise in finance or supply chain do not see themselves as cloud technology providers. Many would happily continue selling Ax 2012, and would be even happier if it came with the browser interface. Sales come via partners many of which are not suited to the new business model..

    And yes I can paint the opposite picture of the benefits of cloud but everyone seems to be copying and pasting the same Microsoft cloud centric mantra so its important the other factors are also considered..

    Like

    Reply

    • Hi. Your comments and views are highly appreciated. Thank you for sharing your thoughts to this blog. The transition to cloud based ERP takes time, but how I see in the marked is that there are now more customers buying into the concept and we have reached a tipping point in trust and maturity of the services Microsoft is providing. For the company I work at, there are no longer any demand among our customers requesting on-prem solutions. This could be a nordic situation…

      Like

      Reply

  4. The whole cloud vs on prem is an interesting challenge. So in my company our profitability is measured before depreciation of assets is taken into account, therefore Capex purchases are far more favourable than an opex spend and therefore Cloud is not the easiest sell internally !!

    Like

    Reply

    • Well… That is not my experience. If they are enclined to not be future proofing and want to waste money on It they could always invest for on-prem. They have an option for a local business data with Dynamics.
      However… Only reason to go LBD is if you have infrastructure already or you’re forced to do so for legal reasons.

      Like

      Reply

  7. “This minimum configuration will estimated be able to support 240-1200 users.”

    Most Ax customers are not that big so this “minimum” is for large customers. For a normal Ax customer i think you can get away with a smaller implementation…..

    Like

    Reply

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.