So it's a question whether they think they need to beat AWS on price or not. AWS pricing certainly makes it very clear that Amazon have no interest or need to compete on price - a lot of the cloud market is extremely price insensitive.
It's a curious effect that seems to be largely down to businesses not used to paying as close attention to the long term opex costs of cloud infrastructure vs capex spend on servers etc.. E.g. I have had clients who paid more just for bandwidth with AWS than their whole infrastructure cost after migrating them elsewhere (AWS bandwidth is easily anywhere from 5x to 50x as expensive as elsewhere depending on location)
I've done the datacenter thing many times and there are so many pieces to the costs of running your own that most people don't include. The biggest one is just plain agility. Predictable workloads can benefit from self-hosted, but then again you can get much better prices at AWS when your loads are predictable as well if you're willing to commit.
- List Prices are cheaper [-1]
- Google Sustained Use Discounts are like RIs but applied automatically and without commitments (up to 30% off if you use a VM for 100% of a month) [0]
- Preemptible VMs are a fixed 70-80% off, rather than the uncertain and highly variable Spot Instance Market [1]
- Custom Machine Types offer you the ability to set your own CPU/RAM [2]
- Both Google and Microsoft offer per-minute pricing
- VM sizing recommendations will give you a message when you can save money[3]
In the end, Google Cloud can be 40% cheaper than AWS[4]
[-1] https://cloud.google.com/compute/pricing
[0] https://cloud.google.com/compute/pricing#sustained_use
[1] https://cloud.google.com/compute/docs/instances/preemptible
[2] https://cloud.google.com/custom-machine-types/
[3] https://cloudplatform.googleblog.com/2016/09/six-Google-Clou...
[4] http://www.zdnet.com/article/what-google-says-to-aws-price-c...
(disc: work @ Google Cloud)
They have a huge suite of offerings, good reputation, huge amount of third party tooling, and the services themselves are somewhat sticky.
By that I mean, I could probably switch over to your cheaper cloud service. It would mean having my engineers go through and build stuff to fill in the gaps or transition away from things that are very AWS specific.
Or I could just pay a bit more money and keep developing features as AWS isn't too expensive and does what I need.
Moving to cheaper services quickly becomes a big deal for people with low margins and high infrastructure costs, though, where moving to cheaper infrastructure can make the difference between being able to match competitors on price or not.
The key is to understand your hosting needs well enough to know whether or not you can afford to build a dependency on AWS specific services.
They will drop their publicly listed prices once smaller customers gets comfortable enough with cloud services to start seriously shopping around, not before. Larger customers will already negotiate (e.g. I know of one AWS customer with ~75% discount over the listed prices)
> The only place I have a gripe on AWS pricing is bandwidth. I keep waiting for the announcement that bandwidth is being cut in half next month and it hasn't happened.
AWS bandwidth pricing is 5x-50x higher than the providers I usually use. E.g. for the cost of 1TB of transfer at AWS you can rent a server for a month + multiple TB of transfer elsewhere. Cutting it in half would still leave them ridiculously expensive for bandwidth.
> I've done the datacenter thing many times and there are so many pieces to the costs of running your own that most people don't include. The biggest one is just plain agility. Predictable workloads can benefit from self-hosted, but then again you can get much better prices at AWS when your loads are predictable as well if you're willing to commit.
I do the datacenter thing every day, and I also do AWS and other cloud providers every day, and when I move customers to dedicated hosting they typically cut their total infrastructure costs to 1/2 or 1/3 vs AWS costs after considering reserved instances and maintenance costs. For many of them, if they were large enough to get serious attention from Amazon they might have been able to negotiate deals that'd make it cost effective to stay, but I've also had more extreme examples, like a customer that cut their hosting cost to 1/10th.
The agility argument doesn't fly with me - I manage multiple systems that are hybrid mixes of dedicated predictable loads on self-owned or leased hardware, coupled with cloud resources that are "tied in" via VPNs as needed.
With well designed systems, the very availability of cloud makes dedicated hosting cheaper even if you never or rarely use it: Where we before often had to plan for 50% to 100% overcapacity on dedicated hardware to handle spikes (and might still have to just accept that some spikes would slow things down), we can now plan on almost no overcapacity and spin up on-demand instances to take spikes (we need only sufficient overcapacity to react and for additional capacity to come online), or even to shave off the top peaks of daily traffic.
And not a single one of them has anything close to the wide and growing offering that AWS presents.
> AWS bandwidth pricing is 5x-50x higher than the providers I usually use. E.g. for the cost of 1TB of transfer at AWS you can rent a server for a month + multiple TB of transfer elsewhere. Cutting it in half would still leave them ridiculously expensive for bandwidth.
But that bandwidth is usually capped in speed and is dedicated to only that instance. No provider like this can handle all of their clients using 80% of their allotted bandwidth. AWS's bandwidth may be expensive, but it is as elastic as you can get and you are never over-provisioning. It's also negotiable just like anything else.
> but I've also had more extreme examples, like a customer that cut their hosting cost to 1/10th.
I'd be interested to hear particulars of this. Obviously it is your edge case, but I find it hard to believe that somebody can cut their costs that much unless they were way over-provisioned on AWS or they are not really counting all of the costs. I have never seen a DC that didn't have more management overhead to run than a full cloud offering.
I don't think that's an accurate statement. Google Cloud and Azure have broad offerings, and in many cases eclipse AWS (ex: Big Data, ML, pricing, BI, NoSQL, Hadoop/Spark, containers, and even things like load balancers and persistent storage).
And that's why it is sometimes worthwhile using AWS. But most people don't use any of the advanced features of AWS.
> But that bandwidth is usually capped in speed and is dedicated to only that instance. No provider like this can handle all of their clients using 80% of their allotted bandwidth
You are jumping to conclusions that I'm only talking about "all inclusive" server rental deals. That's not the case.
You can buy bandwidth at a committed information rate charged based on a 5 minute 95th percentile billing which will typically end up at ~10% of AWS cost with ease unless you have extremely high/lasting spikes in a vast number of data centres from a large number of providers.
Much less if you take advantage of peering or buy limited transit (e.g. if you're in the UK, and most of your traffic is within the UK, you can cut your cost drastically once you reach any kind of volume by getting peering agreements in place for most of your UK bandwidth use, pay a low cost for Europe-only transit, and pay a higher cost for a much lower amount of worldwide transit; it has a management overhead so it doesn't pay until your bandwidth bill starts growing, which is why most people will start off by taking "only" straight unrestricted transit at a higher cost per Mbps/Gbps)
You can also pay per GB/TB and end up paying far less than AWS.
E.g. at Hetzner, you pay per TB over a minimum limit, just like with AWS. For Hetzner, the current overage rate is €1.17 per TB. I have no reason to assume that they're losing on money on that - they'be be able to hand most of that off via peering agreements.
> I'd be interested to hear particulars of this. Obviously it is your edge case, but I find it hard to believe that somebody can cut their costs that much unless they were way over-provisioned on AWS or they are not really counting all of the costs. I have never seen a DC that didn't have more management overhead to run than a full cloud offering.
They moved EC2 instances to Hetzner instances of roughly same speed. That's all there was too it - Hetzner server prices are low, and their bandwidth prices are as mentioned about €1.17 per TB above a limit per instance that's large enough that they've yet to exceed them, and that's the thing that's making the biggest difference (I've actually in the past cut peoples AWS bills by putting boxes at Hetzner and elsewhere purely as caches in front of their AWS setups because of the huge bandwidth differential - for people who spend much on bandwidth, it can pay for the extra servers many times over)
My experience is that we had no problems meeting similar bandwidth levels per machine as they did at AWS. Management is roughly the same complexity. They ended up using "bare metal" dedicated servers (Hetzner also offers VPSs) so they have access to IPMI to manage reboots etc. as necessary, and Hetzners rescue system lets them reimage the servers at will, so after investing a few k on scripting server installs it's as automated as their AWS setup was. They pay a very minimal extra overhead in an encrypted VPN between instances (and with current setup would pay for bandwidth for that if they exceed the included bandwidth allotment), though you can avoid this with some extra complexity (you can pay Hetzner some small fee and get servers wired up to a private network).
The cost difference includes the per-hour costs paid to keep either service running - the latter setup has actually cost them less in time spent managing it.
Manpower and a Stable network is anywhere from 100x to 1000x as expensive as just hosting on AWS.
I've not ever seen any non-AWS setups where manpower or network stability contributed anywhere remotely enough to operations to eat up the cost advantage over hosting elsewhere.
There are good reasons to use AWS, if you can leverage their various services, and hosting isn't a large part of your cost, but price is not it.