[mrweasel]

My first reaction was that you're wrong, but you're probably correct in that large scale VMware installations are much larger than 3000 servers. The small side is going to be a large set of customers, probably the majority.

There are tons of VMware installation that are 3 - 20 physical server, more so than 10.000 server installations. Right now Broadcom is telling those customers, and those with 500, or 1000 servers, that they can take their business elsewhere. Broadcom is much happier to milk the top 5% of customers, compared to servicing the "bottom" 95%.

Broadcom seems to be cutting of to large a chunk of customers. They also are preventing new large customers from coming in. No one starts out with 10.000 physical servers and no one wants to switch a growing 5000 server datacenter from HyperV, KVM or Nutanix to VMware.

[jerf]

In this case you're probably better off thinking of these things in roughly logarithmic terms rather than absolute. By the time you're up to 20,000 servers, you are going to have roughly the same issues to solve as a 10,000 server install and a 30,000 server install. In absolute terms there may be VMWare installations that seem much bigger, but in logarithmic terms there aren't that many more classes above them, and they get rarified quickly. The same ways that you're telling this customer to get stuffed are going to be affecting even the bigger guys in a way that you can't necessarily compare the experience of this customer to the "10 VMs on one server" customers. It's risky to do that. I'm not sure there's as much differentiation in the market to be exploited as Broadcom thinks. Even their larger customers are not just going to sit there and take it... they're just going to move more slowly. But at that scale, VM migrations have a very natural granularity to them; you move them one at a time. You don't need to drop VMWare next quarter; you can just begin your migration project and move away at a pace of your choosing.

But then again, Broadcom may well be explicitly viewing this internally as "we no longer have a competitive product, there is either no way to differentiate it again or we're not willing or able to invest in what it would take, and so the game is just to milk all the money we can out of this", in which case this is pretty close to what I'd expect to see. Where I personally sit in their customer base now, VMs are a commodity and I've got no reason to pay for them anymore. That's been true for a while. I'm speaking here just as A Guy, not as an employee of anyone, but I would expect that this reality has been slowly but surely crawling up the market from where I sit at the bottom.

(This is just discussion, not "disagreement".)