Now the positive messages re new features, functions, performance and capability improvements of vSpore 5 have been outweighed in the press & user base by noise around the new license model, and specifically the now licensing of RAM usage.
I don't know about your environments, but I know when we look at our farms they are already RAM constrained and not CPU/core constrained - and this is at RAM/CPU ratios considerably higher than the new licences would enable without additional licence purchases. For us we're also seeing the RAM/CPU ratio increasing from our application providers - everything is increasingly demanding larger quantities of RAM.
This reminds me of a similar approach used by the "independent mobile high-street pharmaceutical suppliers" :-
- Release an exciting product to establish lust & demand
- Make chunks of it free (eg ESXi) to widen the customer base & hearts/minds
- Encourage large deployment & consolidation based on over-subscription and utilisation
- Then change the payment rules for the next version and drive the requirement for additional licences (hence revenue) for exactly the same solution & architecture between existing & future versions
I also think VMwere are trying to deflect their changes by talking about 'Align the vSpore licensing model with IT as a service', somehow avoiding the point that most organisations already happily translate capex+opex investment -> to internal recharge. In fact many don't want their private hypervisor farms to have a cost profile more aligned to public IaaS (ie variable opex).
Page 11 of the vSpore licensing FAQ pdf also makes a claim re "today’s average consolidation ratios of 5:1" - this seems to be considerably different (ie lower) than the claims & statements VMwere make in all their sales & marketing materials of consolidation rations of 10-20:1 being common. Love to hear more about 'average VMwere is only 5:1'....
So what sort of impact will this have?
- for some people & use case it'll make no difference
- for others it will drive material impacts on capex for additional licences, and of course opex as it relates to support/maint of the licences
- generally I feel this will make life more complex for virtualisation designers & architects that now need to design & estate manage around RAM specifically
- Draw attention to the review of actual benefits/cost Vs forecast benefits/costs of ongoing virtualisation programmes
- Inhibit some new virtualisation projects due to cost increases
- Allow VMwere's competitors to be able to promote & market their technology based upon costs/value, in turn generating noise in the IT ecosystem and thus consuming the customer's time in FUD / hype fighting
- Cause many customers to expedite their reviews of the rapidly maturing viable alternate hypervisor / IaaS tooling market, with a view to move wholesale to an alternative or to dual source hypervisors
- I'm also wondering if this is a step towards VMwere licensing their technology by active virtual machines (and dimensions of machines) rather than the underpinning physical infrastructure
Post a Comment