Amazon AWS: 'Hi there!' VMware: 'We submit. Please, save us' Amazon Web Services and VMware have agreed to work together to make VMware's vSphere server virtualization software available on AWS infrastructure. At a media event held at the Ritz Carlton in San Francisco on Thursday, Andy Jassy, CEO of AWS, said that in recent years, enterprise customers have been confused about the nature …
The press release makes it sound like VMware is trying to get some short-term revenue in a space where they failed to compete from customers that still have brand loyalty.
This partnership is dangerous because Amazon watches everything that happens on their platforms; they are known to make moves on valuable things that are hosted by AWS.
Virtualization technology is mature and the particular flavor of it hasn't mattered for a few years, especially since containerization. I wonder how long VMware will be able to keep premium pricing vice the built-in solutions of Linux and Windows.
Disclaimer - I work for VMware.
I think this post is somewhat inaccurate in it's suggestion that the virtualization layer does not matter and has not done for years. There is some validity to the thought that other virtualization platforms have improved; but at the same time so has the vSphere platform to maintain a significant gap. The VMW virtualization platform provides the most manageable, performant and secure virtualization platform available.
The point about containers is not fully considered on a number of fronts.
1. Containers are not ubiquitous and it will likely be a long time till they are due to the level of fundamental change required to an application.
2. Containers have a place; but still it has varying degrees of value pending the application use case. Therefore the effort and cost associated with re-architecting an application for a container solution in many cases will not make sense.
3. Containers are not mutually exclusive with virtualization - virtualization can provide many complementary features to containers based applications such as greater security, simplified container OS management, stronger resource management to name a few.
4. Many customers I have worked with want the operating cost model of Amazon, and access to the diverse range of AWS services; and many make initial statements of have a desire to migrate into AWS. But the reality is that after an extended period (read years) the majority of these customers have utilised AWS for some net new applications; but much of their estate is still on premise. One of the reasons for that is the degree of complexity in 'transforming' the on premise application and data to run on AWS. This announcement removes that barrier and make it a frictionless ability to move from on premise to off premise and back again.
I think this is a massive win for customers and I think it will be something customers are very interested in exploring further.
Disclaimer-I don't work for VMware but I have a vCloud Air account and it is far behind AWS and Azure and would never have caught up on its own.
I think the main point is vmw was forced to do this. It is plan C, not plan A. You fucked up plan A big time and patty even shot off his mouth on how the vmw cloud was going to rival AWS and Azure... oopsie. Customers might like it ... or not. Time will tell and its a long trip from here and a lot can go wrong when you depend on competitors for your success.
Disclaimer - I've been running VMware as a 24/7 server platform for 13.5 years, and I've just updated my VCP.
For the vast majority of virtualised payloads, the hypervisor doesn't matter. The ecosystem around them does.
VMware has a nice ecosystem from being first to market. They have really easy and really scalable management tools, and have great support from the likes of Veeam. But they charge a significant premium because of the name. Sure, they do have the best features, and if you need *that one feature* that only they offer then you're fairly stuck. On the other hand, for the rest of us the other hypervisors are "good enough", as a rule.
Microsoft is dick-waving with Hyper-V 2016 VM stats, which is mostly irrelevant because the hardware is not there to keep up. However, they also have an integrated management platform (SCVMM, which is a pain in the arse to get set up, but seems pretty decent to me), and Veeam support. For most companies that's really all they need.
It's been said before that VMware is "legacy". They have nice things like e1000 NICs, and can effectively fool the guest OS into believing that it's king of its own castle. And VMware plays that game better than anyone. KVM, Xen, Hyper-V all want to play in a space where the guest cooperates, and is aware that it's running in a digital version of Plato's Cave. For some virtualised workloads that's simply not good enough. But at the same time, how many people still have an old Win2k VM kicking about?
Did I say that VMware was expensive? I'm flirting with Hyper-V / SCVMM simply because we can save half of our VMware licensing costs by ditching VMware and going Microsoft. And I didn't think I'd consider that even 3 years ago.
Agreed this is plan C (if not E or F - losing count). But I thought they were making a play to be the layer that hides multiple clouds (so AWS, Azure) etc. So is that plan disappearing to?
And now they've got a new favourite friend, and is their old best friend (IBM) going to like that. Or will both AWS and IBM disappear when a Dell/EMC cloud appears?
Interesting to see this pitched as "win/lose": I wonder if that's the real story? My perspective (and it's very much an IMHO) is that cloud and virtualisation are rather obviously facets of the same gem, which is the commoditisation of processing: if you have logic that needs to be executed, you go wherever that can be done quickly and reliably. (If you pursue the agnosticism far enough, you wouldn't even care in what language/s the logic was implemented, what tin it runs on or physically where that might be, what OS it depends on or what backend database it uses. Why waste valuable brain cycles worrying about something which should be a black box?) So theoretically it all comes down to price.
If that were true, you'd simply evaluate cloud providers on a bunch of measurable metrics (CPU, storage, bandwidth, uptime, scalability etc), negotiate a price and let the spreadsheet decide who wins.
But while this is fine for the essentially quotidian business of crunching payroll or running the shopping cart,it is NOT fine if you are deploying a key part of your business's unique selling point, handling confidential data, achieving competitive advantage, storing high-security information for third parties. Because in allowing that stuff to run on someone else's black box you are offering hostages to fortune.
To put it in stark terms: if you set up an online sales operation on AWS, offering desirable products via a great website experience, and it becomes really successful—to the point where it's taking some Amazon market share—do you think Amazon won't notice? They are hosting an operation that is cannibalising some of their own sales: given your respective positions, do you expect your business to survive that? We already live in a world where small businesses with new ideas are routinely ambushed and crushed by bigger businesses that "just happen" to have got the same idea and rushed it to market. A world in which NSA, the dominant electronic spying agency, has for decades pillaged the secrets of foreign companies. Where China and Russia devote vast efforts to data theft. Where the world's most valuable company can consider it worthwhile to argue about a rectangle with rounded corners.
Black box or not, the critical question will become: how can I keep my critical, valuable secrets and advantages safe and secure? And while it is possible to encrypt data before you put it on the cloud, that's not much use if it's a 10Pb database that requires constant query.
So it seems to me that cloud will occupy a large place, but sheer survival will dictate that a very large minority of data will always be kept on-prem, under lock and key—and the locks and keys will be better ones than we use today.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
