Amazon CEO says AWS staff now spending ‘much of their time’ optimizing customers’ clouds Amazon Web Services sales and support teams are currently “spending much of their time helping customers optimize their AWS spend so they can better weather this uncertain economy.” So said Amazon.com CEO Andy Jassy in his annual letter to shareholders published last week. “In AWS, like all our businesses, we’re not trying to …
Or, to cut through the corporate bullshit.... (please replace the cloud-provider with the any of the others - it's still relevant for them all!)
Amazon:
"We recognise that the utopia of low-cost we sold to customers was a myth. Our customers have recognised they are spending 2x/3x what they did when they self-hosted.
We are absolutely desperate to retain our customers, before they go back to what they used to do"
Customer:
"We now recognise that we were sold a dud/lemon. We now recognise that maintaining the cloud infrastructure still requires us to pay staff to maintain it. We now recognise that our original pre-cloud setup was cheap!! "
99.9% of customers....
The end.
Not necessarily...
I have a friend who works at AWS, and his role is specifically to optimise customers AWS usage. As he puts it, his target is to save customers money.
The first customer he worked with had done a typical lift and shift of an entire server into AWS EC2. After just a few hours of reviewing what they were doing he was able to propose changes that would get their annual costs from in the multiple thousands down to $100/year for that one server. And this hasn't been just a one off for him.
There is still a lot of naivety in the minds of people (perhaps within management more?) about how to move from on prem into the cloud. Lift and shift as a phase 1 is usually how they start, and then the inevitable happens and phase 2 never happens to actually benefit from potential cost savings. Then the surprise at lack of expected savings.
Of course everyone's experiences will be different, so YMMV.
"I have a friend..."
I don't believe a word your friend told you.
"...benefit from potential cost savings."
There is no cost savings for at least 99.99% of use cases as that's not what cloud is really about. However, that is what has been sold to people and now companies are trying to negate their false advertising statements by laying the blame solely on the customers ignorance. Cloud is supposed to be about ease of use but, there's much debate on that topic.
There's this thing called Jevons Paradox. If AWS tech you to use AWS in a more cost effective way, then the effective per unit cost of doing a thing in AWS goes down for you. At a lower per-unit cost, it becomes economical for you to do more things in AWS. A lot more. So your total spending goes up and you are happier about the results.
It's mainly known for being the reason why you can't fix road congestion by building more roads
This is an honest take. Does AWS cost more to an SMB than hosting their own kit? Absolutely. But...
1. No SMB I've worked for has been keen on paying for multi data-center redundancy.
2. Scaling back compute resources (standing up a server temporarily or getting rid of a server you don't need anymore) is a lot easier to do in the cloud than for an on-premise server for which you are capitalizing for years.
3. On-premise equipment often is expected well-past its expected life span
A lot of the problem with AWS is that many use it as a VM farm with redundancy and stop there. If you don't learn how to *properly* manage auto-scaling and avail of stuff like serverless technology, you will spend a lot more to do the same thing, albeit with more redundancy and reliability.
It's about manipulation, so the friend would be telling the truth: That their job is to help customers reduce their spend.
However, that doesn't make it cheaper than self hosting - it just makes it cheaper than what AWS sold them or encouraged them to do in the first place.
So step 1 is to get the customer to move to the cloud by promising whatever the customer is likely to believe. Then the price starts going up, so AWS (or other cloud service) offers to help reduce cost. They do so, but they only need to reduce the cost to the point the customer sees migrating back to on-prem as not being worth the hassle in the short term. Then, once the customer is again happy, prices go up.... and the cycle of cost cutting / price increase continues so the cloud provider is able to milk the customer for longer as the customer is being manipulated into believing they're getting value for money.
As with anything: These services aren't there for the benefit of the customer: They exist to make the service provider money and anything they can do to fool the customer into believing they're better off is worth doing.
Until something happens and the cloud provider turns off the servers or ejects the VM's and puts the customer out of business. Oh - that's one they don't mention, do they? That if the provider makes a mistake... that could be the end of the cloud service provision for the customer (hey, our monitoring system detected an anomaly with your VM so has isolated it. Sorry that it's your main production server and you now don't have an online presence, are loosing millions a day and this is trashing your rep - we'll provide an engineer to help fix things, as long as you have this premium care package. You do have that, don't you? No? Well, let's set that up for you right away...)
I've oft-compared using the cloud to be much like using takeaways and restaurants vs. cooking for yourself.
Sure, it's easier, you don't have to know how to cook, you can be lazy about it (up to a point), talented experts can come up with things you never imagined, you have outsourced your energy costs, and so on. But you pay a premium for it and are fully reliant on others to make it all work.
Not necessarily, at a previous firm we had a service that created office docs - ppt, pdf, xls, doc - and occasionally it had high demand. To support that demand, on-prem it had 15 beefy VMs and a Ceph cluster. If you lift-n-shifted that to cloud, it would be bloody expensive. Replacing it with lambdas meant that it had more scalability than the on-prem - we could handle thousands of simultaneous doc generation tasks versus tens before - and in the quiet times cost us virtually nothing.
"The lazy naïve way to "lift and shift" from on-prem to the cloud is to create a new EC2 for each existing VM. That never works out, cost-wise."
It depends. I was going to put a complex answer to back that up, but I'm already into pages of ifs and buts. Some short bits - if you are already in a managed DC (even if it is one you own), the cloud is probably cheaper than your current vendor. If you have strong DR requirements, the cloud is probably cheaper. If you are already clustering your workflows using Openshift or K8s, the cloud is probably cheaper).
The implementation details really, really matter though. As you say, life and shift is only step 1...
“Many of these AWS customers tell us that they’re not cost-cutting as much as cost-optimizing so they can take their resources and apply them to emerging and inventive new customer experiences they’re planning,”
But will AWS's customers' customers appreciate the inventive new experiences?
As an Amazon (not AWS) customer I'd very much like to see some optimisation of their search engine so that it delivers hits that match the search terms and nothing else.
I fear, however, that what I actually get is what was once a new, inventive customer experience that Amazon's marketing thought would be ideal - it's not a bug, it's a feature. Never underestimate the ability of marketing to screw things up for customer.
Just finished a project, where we created an internal website for knowledge management and some nice tools, micro service architecture. Internal cost for hosting the previous system internally and the new one in AWS are quite similar. But with AWS getting 8x computing resources during peaks last 1min, vs. the old scenario 3 month. Security is better, monitoring is way better, customer support great. The assortment of PaaS services amazing. From my POV it's not cheaper but better.
Desktops/laptops.
On-prem networking and internet uplinks.
Software licenses.
These are all on-prem costs and almost every company has them, so there's a baseline which can never go into the cloud unless everyone works from home and does BYOD.
There's also significant legacy on prem infrastructure which needs maintained, supported and occasionally upgraded; most large companies will still be running those for a few years, despite any cloud aspirations.
That said, I'd have thought the ratio would have been closer to 70 or 80% on prem.
Hold back the beancounters, but cost isn't the only way to measure success. Can I host a website on a RaspberryPi in the corner of my Home Office - yes. Can I host the same site in the cloud for a few pennies a month and get access to Enterprise grade availability, security, scalability etc - also yes.
And thinking of the cloud as a drop-in replacement for a data centre and expecting it to magically be cheaper is naive at best. Re-architecting to take advanatage of the cloud services can have a big effect on your success KPIs, availability, reliability, scalability etc. Even cost. Work 8-6? Turn your data centre off outside of those times to save costs ;)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
