vSphere5 licensing – the good, the bad, and the ugly

The announcement on 12th July about vSphere5 was largely overshadowed by the furore around licensing changes. My gut reaction was much like many people – angry that VMware seemed to be charging more for the same functionality. If you want a feel for customer feedback, this VMware communities thread is a good place to start or see how many posts on the ESXi v5 forums relate to licensing. I’ve now reached phase 5 of ‘the LonelySysAdmin’s 5 stages of VMware licensing grief‘ – acceptance.

The Good

  • I’ve done the maths for my environment (thanks to Hugo Peters for the PowerCLI script to check) and I’m one of the 90% that VMware claim will see no increase in costs. We’re using about 62% of our vRAM entitlement (using 2.1TB from 3.4TB allowable) so have some growth factored in. So far, so good and not a big surprise as I knew we didn’t push our current infrastructure too hard.
  • At the recent London VM user group there was a similar feeling – many people were OK with the licensing today but had concerns about the future.
  • There are no longer any restrictions on number of cores per socket. My company use Enterprise rather than Enterprise+ so without this change we’d be restricted to six cores per socket, a limit we’ve already reached.
  • Service providers aren’t affected by the recent changes. They’re already on a different licensing model which isn’t based on vRAM (the VMware Service Provider Program)
  • New VDI users can use the vSphere Desktop edition which doesn’t include the vRAM based license model. Our company haven’t gone down the VDI route yet, so we’re not impacted by the upgrade issues (see below).

The Bad

  • Where does this leave the ‘virtualisation’ blades, and systems like UCS which are built around memory density? Useful for service providers but for traditional customers their cost effectiveness is much reduced.
  • What about the existing customer base who have bought into the ‘eggs in a basket’ marketing from VMware and scaled up? The costs for running a dual socket server with 384+GB of RAM has just doubled if you continue with VMware.
  • If you’re a current VDI user (using a product other than View) you can’t upgrade your vSphere licences to vSphere Desktop edition so you’ll have to fit your desktop images into your vRAM entitlement. You can read more at Brian Madden’s blog.
  • What about TPS, http://premier-pharmacy.com/product-category/mans-health/ memory compression etc, features which VMware have repeatedly claimed make their solution cost effective against the competition? These features let you use more RAM than is physically available but when you’re paying for vRAM this is now of little benefit. Where it might be useful is for service providers and VDI deployments (View or XenDesktop etc) where VM density is king.

The Ugly

  • RAM density tends to double in density every two years (according to Moore’s law, which roughly still holds). Will vRAM licensing keep pace? If not your VMware estate is getting more expensive day by day.
  • One of VMware’s initial use cases was consolidation which was a ‘green’ technology, allowing companies to cut heating, power and space. Now that the scale up model is being financially disincentivised we’ll see more servers, more power etc. Not cool VMware!
  • Trust. While every vendor has to have a revenue stream, I don’t thinkg VMware have done a good job of communicating this and previous licensing changes. As the person who has to justify the choice of VMware for my company I’m increasingly wary of promising things to management which, through no fault of my own, I have to retract further down the line. Oracle, Microsoft, Apple – all have forms of restrictive licensing so VMware certainly aren’t unique, but they’re gaining a reputation which previously was more associated with their competition. Like all companies we do annual budget planning and changes like this don’t go down well with the budget holders.

All said and done I’m relatively sanguine about the changes. Are they a good thing for customers as VMware claim? Not especially. Is that a consequence of no significant competition and the increasing clout of a large vendor? Absolutely. Paying dues (be that software licencing, rent, taxes etc) is part of doing business and always will be. At least there are alternatives – if you really don’t like it, maybe it’s time to evaluate a change. In the VMware community podcast #76 about virtualising Oracle the VMware line was ‘if you keep paying Oracle, they’ll keep the licensing and support stance as it is – it’s not in their interests to change it’. The same applies to VMware.

VMware are holding some webcasts about licencing on the 18th July, 25th July, and 18th August – tune in if you want to know more.

8 thoughts on “vSphere5 licensing – the good, the bad, and the ugly

  1. Is this really a disincentive to scale up? To run a certain number of VMs on more or fewer physical hosts will require a very similar vSphere entitlement. There can still be saving with scale up, especially guest OS licenses — they are typically done by physical socket.

    Eric [VMware]

    1. Thanks for the comment Eric, good to know I’m not talking to myself! As you rightly point out some licensing is per socket and these may well be cost effective to virtualise but the fact remains that it’ll be more expensive to scale up with vSphere5 than previous versions. Hence it’s still a disincentive, even if the degree is variable.

  2. Ed, thanks for the article, interesting point on the VSSP, I subconsciously assumed this would be adjusted again as well although it is of course just a relatively recent change …
    And Eric, good question … is it really disincentive?
    My take is quite pragmatic … vRAM wont render ‘scale up’ obsolete but clearly impacts the value proposition, especially for systems with memory extensions. Proposition was tangible and simple… reduce the average $/vm by increasing the cons ratio through memory for the same license cost …thats now mostly gone. Are there other +’s..? Sure, fewer managed end points etc …but it will be harder (not pointless) to make a cost case given that those boxes typically have a higher entry point to start with… You really need clients who understand and value the ‘softer’ tco benefits and are very comfortable with the ‘many eggs…’ approach. Just my view …
    Andy

  3. Customers will give minimum vRAM to VMs. There will be lot of paging. Performances will decrease. vSphere5 will be the worst plateform ever.

    1. Thanks for taking the time to comment Cedric. I disagree with your statement about vSphere5 being ‘the worst platform ever’ as if used correctly performance will match or exceed the current platform which is widely acknowledged to be the best available.
      I do however think you’ve got a point about the risk of people allocating a minimum of vRAM and therefore affecting performance. ‘Right sizing’ a VM has often been talked about but in my experience less commonly done – it’s not always as straighforward as people think and the incentive wasn’t as strong previously. We use a large number of Java VMs and Oracle VMs which are harder to right size as the JVM or Oracle SRA handle memory allocations instead of the guest OS. Maybe ‘right sizing’ will get more column inches as vSphere5 starts to permeate the ecosystem?

Leave a Reply

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