I took a minute to read the licensing guide for vSphere 5 and I'm still trying to pull my jaw off the floor. VMware has completely screwed their customers this time. Why?
What I used to be able to do with 2 CPU licenses now takes 4. Incredible.
BL460c G7 with 2 sockets and 192G of memory = 2 vSphere Enterprise Plus licenses
DL585 G7 with 4 sockets and 256G of memory = 4 vSphere Enterprise Plus licenses
BL460c G7 with 2 sockets and 192G of memory = 4 vSphere Enterprise Plus licenses
BL585 G7 with 4 sockets and 256G of memory = 6 vSphere Enterprise Plus licenses
So it's almost as if VMware is putting a penalty on density and encouraging users to buy hardware with more sockets rather than less.
I get that the vRAM entitlements are for what you use, not necessarily what you have, but who buys memory and doesn't use it?
Forget the hoopla about a VM with 1 TB of memory. Who in their right mind would deploy that using the new license model? It would take 22 licenses to accommodate! You could go out and buy the physical box for way less than that today, from any hardware vendor.
Anyone else completely shocked by this move?
hmtk1976 wrote:Just buy x GB of RAM worth from whatever edition you want, regardless of the number of CPUs. I
Look at it this way: they did do it just like that.
If you have an existing environment, you already have all the CPU licenses you need, forget about CPU licenses.
If you have Standard edition, you buy additional vRAM in 32GB increments at a cost per GB of Standard$ divided by 32GB,
once you require the full 32gb.
If you have Enterprise Edition you buy additional vRAM in 64GB increments, at a cost per GB of Enterprise SKU cost$ divided by 64.
If you have Enterprise+ Edition you buy additional vRAM in 96GB increments.
The amount of RAM included in each edition is the minimum increment you can purchase vRAM at.
Once you've finished buying enough vRAM, you are pretty certain to have a sufficient number of CPU licenses,
unless your workload is highly unusual, or you aren't picking servers actually suitable for virtualization....
for example, if you're buying servers that have fewer than 6 cores per CPU, then yeah, in that unlikely case, you might run out of CPU first.
And what if I want a 4 host (dual sockets) cluster that's heavy on CPU and light on RAM and I want to use Enterprise or Enterprise Plus licenses? I'd still need to buy 8 CPU's worth of licenses and possibly be left with TOO MUCH vRAM. Of course you can start blabbing about vRAM pools but I'm not interested in that. Fact of the matter is that VMware isn't simply licensing RAM but they continue the per CPU route. They simply want all the cakes in the bakery. In the long run they may be left with stale bread.
Just go to what I mentioned earlier. Devise a verifiable computing units formula, and tell people your logic based on historical research. When you hit the limit of what you bought, things slow down, and it records when and for how long. Then it doesn't matter if you run it on 5 fast computers or 50 slow ones. When technology advances, it doesn't matter where or by how much. They could reduce the cost per computing unit to entice people to move to the next version by having the computing units cost less. By syncing to the normal lowering cost per computing units that advances in technology brings, they maintain user's cost-benefit ratios, and keep people on their gravy train. It not only benefits VMware's bottom line, it adds predictability and stabilization to company profits, which increases the value of the company and its credit-worthiness. Their customers can backbill their customers based on usage. Their crazy licensing by the VM that they have with server vendors would go away and people might actually use it. For VMware, each new version becomes a boardroom calculation based on advances in technology. That number gets plugged into the next version. Customers would be crazy not to upgrade to the latest version because the cheaper cost per computing unit would be with the newer product. The people getting the most benefit, pay the most, and can afford to. The ones getting very little benefit, can still afford to use their product.
What they have is a system to determine the weight of a fish by multiplying the diameter times a constant. Just weigh the fish. Then it doesn't matter if the fish is long and skinny or short and fat.
This way it may make sense. For example, I need 32 GB of ADVANCED vRAM and 192 GB of STANDARD VRAM. SO I can pay for only 32 GB of ADVANCED (or ENTERPRICE, you call it) licenses because my VM-s which require ADVANCED features use only 32 GB of vRAM; and I pay for 192 GB of standard vRAM because my other VM-s need it (but not need ADAVNCED).
But they did not do it so the licensing schema in vmware5 combine both limitations of the previous schema and add new limitations of the new licensing. If I run 4 VM hosts but need some features for 2 VM-s only (total of 32 GB vRAM) I can't purchase 1 ADVANCED or ENTERPRISE license and 6 STANDART - I must purchase at least 4 advanced/enterprise...
VMWARE 6 will come with vRAM and vCPU and vDISK licensing model so you will have limitations in everything. And they will have a very good explanation why these limitatioons means really increase.
VMware 6 free will have 32 GB of VRAM and 16 virtual CPU and 1024 GB of virtual disks limit for sure.
PS. Free VMware ESXi 3.5 is THE ONLY reason why we have now purchased more then xxx standard and enterprise licenses and 2 or 3 essential-s.
If we could not deploy development on free version, we did never go with VMware - we could select free XEN then switch to commercial XEN, or use something else but not VMware. If it was version 5 with it's limitations, it should not happen .
And what if I want a 4 host (dual sockets) cluster that's heavy on CPU and light on RAM
If you need to be CPU heavy, use a 16-core or 12-core CPU deployment.
Dual socket servers that suck more power per unit of compute workload are history, for virtualization,
unless you actually require more than 16 cores pe rserver.
In that case... yes.. that's definitely a very high end workload, and VMware is
quite justified in charging more for dual and quad socket servers that therefore
host 24, 32, 48, or 64 CPU cores.
In any case, you are no worse off than with vSphere 4, and due to the removal of CPU core limits,
you are now much better off, in regards to CPU compute power licensing.
vSphere 4 was idiotic in it's CPU core licensing. Period.
What VMware is "justified" to do is something customers will decide. They're making something that's quickly becoming a commodity more expensive. Doesn't sound smart IMO.
VMware 6 free will have 32 GB of VRAM and 16 virtual CPU and 1024 GB of virtual disks limit for sure.
Yeah I think you hit the point, and qualified for CFO at VMware.
But i think we won't see this before VSphere 7.
I am sure the company will be bankrupt and trying to get a Gov’t bailout by then due to losing a large % of its once loyal customers. Any CFO should no that no customer = no money. The fact that they would make a boneheaded decision to T off so many of their customers will bite them hard I hope.
I am currently working part time and am in the office Monday, Wednesday and Thursday. If you require urgent I.T. assistance please contact the IT Service Desk on ext 6999.
They are like screwing up the virtualization objective thats is first of all = savings!!!!
I agree. As a CTO, I tried to make a case for virtualization for a very long time. Until relatively recently, the licensing was not cost effective. Hardware just doesn't cost that much, there is no new learning curve, and you don't have all of your eggs in one basket. VMware's licensing didn't work except in unusual circumstances. It wasn't until you could specify the virtual CPUs and cores that it made sense for the CPU-based apps and operating systems. It may have been due to a couple feints by CITRIX and Microsoft, but VMware changed their licensing to make virtualization cost effective. As one would expect, the market took off in a world where they had no competition. Afterwards they realized they had no competition. Customers found the virtualization has more value than than simply saving hardware. Now VMware is doing what Mircrosoft has done with Windows in the face of no competition, raise prices. If you were a VMware stock holder, you would expect that. That's why you put your money there. The only way I see this changing is if they have competition. Microsoft has stated that they have no intention of raising their prices with Hyper-V. Of couse not, they can't with their product's shortcomings, plus they use it as bait to sell Windows Server, which they continue to raise the price on because they have no competition there. From VMware's perspective, the worse times get, the less IT can play the "My favorite vendor" game. In theory, the more money VMware makes now, the better chance it has of keeping their competitor wanabes, wanabes. The only fly in the ointment is EMC. If the increased revenue is used to float EMC, then it gives their wanabe competitors an opportunity to graduate. My crystal ball says virtualization will be part of the server BIOS in the future, and built into chipsets, so VMware needs to make hay while the sun shines.
A usage-based licensing model broadens their base by insuring that the cost of virtualization remains tied to the benefits derived. That eliminates current dead zones where their licensing allows users to skate by too cheaply, and where their licensing causes them to not be cost effective. It also means they don't have to overcharge their best customers to make up for it, which can easily cause them to lose the war. They have no cash cow to hook it to like Microsoft does. VMware needs to understand that the ratio of savings cannot be near what customers are currently saving, because their customers are using it for things they would not do at all if it were not for virtualization. Making VMware cost about 25% of the benefit customers derive from it would maintain broad appeal. There is WAY more money in hamburgers than there is steak. 0% of the perspective customers are influenced by the salesmen's fuzzy math when he gives examples of how a bank had hundreds of PCs to call into and he saved them from building larger facilities, and reduced their electric bill. When I was CTO, and a salesman used that on me I was thinking, "That is one stupid bank and the saleman must think I'm stupid. The bank should have been on CITRIX or Terminal Services." VMware needs to base its value proposition on cutting capital expenditures or monthly leases of servers by well over half to get anyone's attention. Customers know there is an average of 25% performance cost to virtualization, and that it takes a pretty obvious and large savings before you can add a technology, and have it translate into real dollars on the bottom line. Even then It has to pay for itself pretty quickly because it has to compete with the ROI of other projects that grow their business or save employee costs.
Besten Dank für Ihre Nachricht.
Ich bin bis am Mo 19.12.2011 abwesend. Emails werden aus diesem Grund erst verzögert beantwortet. Bei dringlichen Anliegen bitte ich Sie meinen Stellvertreter (Fabian Dobler, fabian DOT dobler AT travel DOT ch) oder meinen Vorgesetzten (Pascal Scandola, pascal DOT scandola AT travel DOT ch) zu kontaktieren.
Herzlichen Dank für Ihr Verständnis
Freundliche Grüsse / with best regards / cordialement
>Until relatively recently, the licensing was not cost effective.
>It may have been due to a couple feints by CITRIX and Microsoft, but VMware changed their licensing to make virtualization cost effective.
What licensing was not cost effective and what was the change to make it cost effective?
ESX (1.x/2.x) used to cost as much as a new server, meaning as long as you went 3:1 consolidation your initial investment was a wash. Figuring in HVAC, rack space etc vs training costs and lack of app vendor support for virtualized servers complicated the math from there but the basics were sound.
The major changes I can think of were when SMP came included in all licenses (3.0 I think) and when vMotion was pulled down to Ess+ (4.1). What other "major" pricing changes were there and why do you think it was influenced by MS/Citrix?
>Hardware just doesn't cost that much, there is no new learning curve, and you don't have all of your eggs in one basket.
Unless you are buying a hardware-clustered box your eggs are certainly in one basket with hardware - something VMware has *not* been since 2003 and the release of vMotion.
>It wasn't until you could specify the virtual CPUs and cores that it made sense for the CPU-based apps and operating systems.
SMP (vCPUs > 1) came out in 2003. Adding cores from the GUI came out with v5. So when did the math start working out? And can you give a practical example of the benefits of setting cores per CPU?
>My crystal ball says virtualization will be part of the server BIOS in the future, and built into chipsets, so VMware needs to make hay while the sun shines.
This has been apparant for awhile and explainsVMware's push into management and the explosion of their product line. Mail server, database server, security solution.
>If the increased revenue is used to float EMC
EMC 4th quarter 2010 "consolidated revenue was $4.9 billion,"
VMware 4th quarter 2010 "pulled $835.7 million in revenue"
Who's floating who?
>A usage-based licensing model broadens
They are not using a "usage-based" model but an allocation-based model with loopholes (max 96GB per VM) and caps (192GB for ESS/ESS+).
>Making VMware cost about 25% of the benefit customers derive from it would maintain broad appeal.
The sales pitch should be the value of the added features which would be difficult or expensive to add to existing infrastructure. HA, snapshots, vMotion, DRS/DPM, DR replication (req SRM5), thin provisioning and the flexibility of VMs.
>When I was CTO, and a a salesman used that on me I was thinking, "That is one stupid bank. They should have been on CITRIX or Terminal Services."
Not everything runs on Citrix or TS, which hopefully the "stupid" bank was aware of.
Tho my first View client started off the project with "We will never do Citrix or TS ever again" and has loved their 120 seat View install.