Compared that to a company like Google that can't make any money at all off their infrastructure except with search ads :-) That is the real story for me. And a tiny bit personal because when I was there I didn't feel that platforms got nearly the respect they deserved for keeping the roads rolling.
Congratulations to the Amazon team, that is an impressive feat indeed.
> “Amazon Web Services is a $5 billion business and still growing fast -- in fact it’s accelerating,”
I take that to mean that the business is growing but the rate of growth is also growing (hence "accelerating"; I like how he puts it so succinctly).
Revenue 2013: $1000
Revenue 2014: $1690 (+ $690)
Revenue 2015: $2518 (+ $828)
With those figures you could still argue that growth (in terms of extra revenue per year) is accelerating.
- Q1 2015 revenue: $1.57B (AWS) vs ~$1.5B (CRM)
- Q1 2014 revenue: $1.05B vs $1.23B
- Q1 YoY change: 49% vs ~22%
- Q1 2015 operating income: 16.9% vs (~3%)
- Q1 2014 operating income: 23.3% vs (4.5%)
I'm not taking anything away from Salesforce, it just happens to be a similarly sized comparison that most people thinks is a pretty good business. When compared, AWS looks pretty great!
i don't think AMZN or Google put 4 sockets per 1 U. While obviously it is possible, it would be pretty dense for normally powered chips. My understanding is that 2 socket/1U is more typical. At 100W/chip it would be a difference between 8KW and 16KW heat from a rack. But anyway, you're right - AMZN probably gets more than $1/1U. Their cluster 2-socket (16 real, 32 v-cores) instances go for $3.5/hour. Starting from that number we can adjust it down somewhat for pre-paid/reserved/spot pricing. Anyway, lets hope fuzziness in one number is compensated (not amplified :) by fuzziness in another ... Actually from the 2013 link you posted in the response below (that one - https://storageservers.wordpress.com/2013/07/17/facts-and-st...):
"Amazon data center- Amazon has around 450,000 servers in its data centers sited in 7 locations in the world. "
So adjusting for 1+ years of growth (of at least 10-20% as they had 49% revenue increase) my original estimate was less than 2 times wrong :)
Edit: actually adding storage servers - may be up to 4 times wrong if they have all 1U servers, so the most probable parameter to adjust would be $/1U to something like $2+/1U.
http://www.intel.co.jp/content/dam/www/public/us/en/document... covers it fairly well.
It usually isn't worth going to a mega dense rack (25kW or whatever), because unless you built the entire site with this in mind, you'll probably run out of power before you run out of space. Then you spent extra on these extreme density solutions for no real return. It can still make sense, if the entire solution is thought out.
>> SuperMicro sells a 2U machine with 4 full sized computers in it with 6 drives.
>You can put 12 cores in that
In fact, the quad node contains 4 motherboards. Each motherboard supports 2 processors. So that's 8 processors in 2U, or 4 processors per U.
Each E5-2600 processor could support 8 'real' cores, or 16 Amazon Virtual cores. 4 * 16 = 64 Amazon cores in 1RU.
I agree that our math won't be right, and I think you're in the ballpark. But, the densities can be 'hella' higher than one things with modern gear. It's only 70W for the 8 core processor at 1.8ghz, 135W for the 2.9Ghz version, so if we do 4 processors per rack unit, that's "only" 540W per rack unit. 60 Rack units is 32KW/RU - which is a lot but not impossible. I'm gently simplifying by assuming 0 watts for the rest of the chassis, which is approximately reasonable if you're building a CPU farm.
Is this really true? I know that AWS offers many more services than competitors currently, but for the services for which there is a competing product, is the AWS offering really that much better?
Yes, Amazon offers a bigger variety of products, but the newer ones just scream vendor lock-in without much-added benefit. Why would I use Opsworks if Chef or Ansible work just as well?
Choose anything, except maybe object storage, and there is somebody out there who does it better than Amazon, there is also sombody who does it cheaper than Amazon.
Except for S3, AWS isn't the best at anything, nor the cheapest, but, they are the only ones doing it all, and typically the ones inovating new services. Nobody has all the services AWS has, or even as well integrated, or as granular etc...
If you have very specific needs, then go pick and choose from different vendors (assuming you're ok with multi-site latency), but if you're starting a company and have generic needs that aren't well defined yet, AWS is probably your best bet to start with. People do that, and once their invested so much time in learning the environement, they stick to it.
AWS has been diligently moving up the stack, creating higher level services that are "good enough" for many of their customers, who are increasingly enterprises teams who want to outsource as much as possible to Amazon. They can go all-in with AWS and expect a constant stream of new features and services at relatively competitive prices.
With Google, their cloud platform still feels non-core to their business and they have a history of retiring services when their priorities change. I hope Google and Microsoft become fully viable alternatives, but for now, enterprises seem more comfortable betting on AWS.
My personal experience has been exactly the opposite - AWS is light years ahead of Google Cloud. Does Google Cloud provide Machine Learning service?
For the longest time, it was only supporting Java and Python SDKs. And the Google's API is still very much lacking in terms of features and stacks. In fact the better support and much comprehensive services presents more choices rather than "vender lock-in". Keep in mind, all the auxiliary services are simply the open source technologies builds on top of AWS EC2 - if you like to spend time and money, you can totally build your own!
And this is common, as we've just seen people complain about Google Cloud 4 days ago.
Yes they do. 
> And this is common, as we've just seen people complain about Google Cloud 4 days ago.
To me, this is an isolated incident. On the other hand, seemingly every AWS customer of non-trivial size had to reboot a significant part of their machines in the last year because of Xen vulnerabilities (with no way to predict if you will land on a patched machine).
> To me, this is an isolated incident.
Apart from that, noisy neighbors happen on any cloud platform, and as the author of the article said, those test were far from scientific.
I personally don't know why one would use any of the GCE/AWS/Azure ML offerings apart for really small use cases (in which case you probably don't have enough data). Most businesses with an ML component (like my startup) need millions of predictions per day which would be crazy expensive on any of those platforms.
You are describing Google App Engine circa 2010. I recommend taking another critical look at what Google has available because it's quite competitive these days :)
I do like Google's new DB. It is like an easy to use spark w/ cassandra backend, all behind the scenes. Very nice.
Doesn't Amazon's mysql api fronted db work similar to google's?
For scalability, RDS gives you up to 3TB storage (1TB for SQL Server, 64TB for Aurora) and DB instances with up to 244GB RAM, 32 vCPUs and 10Gigabit network, whereas it looks like Google Cloud SQL is 100GB and 16GB of RAM. It looks like Cloud SQL provides managed replication/failover as standard; with AWS this is optional, but simple to set up. (I use RDS; never tried Google Cloud SQL.)
(I have no connection with either company, beyond being a customer.)
In all fairness AWS had a 4-5 year head start on Azure. Also there are a lot of really exciting things coming down the line for Azure. I don't expect this gap to last long.
Competition is good!
But no TLS termination? At a company that has invested $15bn on their cloud products and has in-house expertise on the actual source code of one of the most widely-used networking stacks? That's... surprising, to say the least.
> An SSL termination proxy is a proxy server that is used by an institution to handle incoming SSL connections, decrypting the SSL and passing on the unencrypted request to the institution's other servers (it is assumed that the institution's own network is secure so the user's session data does not need to be encrypted on that part of the link). SSL termination proxies are used to reduce the load on the main servers by offloading the cryptographic processing to another machine, and to support servers that do not support SSL, like Varnish.
- Azure Cloud Services (PaaS)
- Azure Websites (PaaS light)
Unfortunately, when I last looked at it four months ago, both of them were quite unattractive, at least for Python hosting.
First of all, they have no support for automatically installing python dependencies, i.e. no support for pip. If you want to install e.g. django-rest-framework, you have to install it locally and then manually copy the installed files to Azure (with FTP, if I remember correctly).
Secondly, they both run Python on a Windows Server running behind the IIS web server. Besides the fact that you will walk along the road less travelled (Django is overwhelmingly deployed on Linux), there is a very concrete limitation: if you want to install binary packages (e.g. PIL or Pillow), you have to manually compile them for Windows (and of course copy them manually to your Azure server). Yuck.
If you want to use Azure, go with their IaaS offering (Azure Virtual Machines) and stay away from their PaaS offerings
We still used Amazon Email and CloudSearch though. Azure has no email and Azure search is too new for me to trust.
That said, I was able to setup a quick little proof of concept web/worker app with node pretty quickly... though you should be aware it is running in a windows host (either 32 or 64bit configurable)... you can include exe/dll's in your package as needed too... I used a git endpoint for deploy, and that worked pretty well. There's apparently a build environment suitable and configured for node_gyp to be able to use.
Though imho it's not as flushed out as App Engine, or bits of AWS... I am able to run dokku on a few hosts... Work's account finally came online for the SSD preview about a month ago... it's definitely worth it.
There's some mishmash of availability... I kind of like that using blob storage, tables and azure queues is as simple as having a storage account... though the node client is a little lacking in documentation and examples.
Don't get me started on the Beta UI vs normal UI. Some things are only supported in one and not the other and vice versa. I'm constantly being flipped back and forth between the two. Also they are both really unpleasant to use. I get lost a lot.
On the flip side, there were some issues with connecting to Azure via the node `mssql/tedious` module, which isn't from MS, and a few developers from MS Azure SQL teams stepped up to track down the issue. Learned a bit more about the TDS protocol than I really wanted to that month.
It seems to me that MS Azure's development department should either implement a 20-40% time for developers to work on the public stuff and/or documentation as developers desire, or have cowboy teams that do. MS isn't likely to spur more open development efforts as most corporate based projects don't inspire much community involvement.
That said, given the choice I might be more inclined to use Joyent, DigitalOcean or Linode for my own stuff... The choice wasn't mine at work, the existing solutions are mostly .Net (some Node) based and Azure was the most pragmatic choice of migration to a hosted solution when that was done almost two years ago. And running anything under Ubuntu VMs with docker on Azure is pretty trivial... Once the next generation infrastracture is in place (a couple months away), it will become easy enough to run anywhere.
They could have summed up the whole thing as First-mover advantage. Amazon created this new market and when MS, Google etc saw a real opportunity in it, they wanted a piece of it and thus moved in by replicating the best features. Meanwhile for Amazon it was all about maintaining its majority share refining and expanding.
Depends on your needs, tbh, traditional hosting is a better fit for mine. I can automate buying and provisioning bare metal servers already. I don't need AWS for that.
Yes, AWS is way ahead of the other cloud providers in every way.
BUT scale is not the only thing that the cloud brings. AWS is not just S3 and EC2. I think it's pretty clear that Amazon is much further along the road of productizing their cloud. i.e. providing all the products that businesses and most notably enterprises need to move to the cloud. It will take a while for Google (and Microsoft) to catch up in this area.
My experience with Kubernetes and Managed VMs and as an App Engine dev since 2009 compared to Amazon's latest offerings such as Glacier and RedShift shows Amazon is distinctly targeting the present with these legacy plays. Google is targeting the future.
Not much talked about: Google's profitability gives it a distinct advantage in its ability to wage a price war compared to Amazon should have anyone on Amazon's platform concerned.
In other words, Google is building a cloud that you can only really pick up for blue-ocean ventures. AWS gives you a bunch of the same tech, but also a clear transition path: you can first duplicate all your present-day infrastructure in their cloud with little-to-no service interruption, and then start replacing it, component-wise, with the future-tech.
"We're not going anywhere. We're NOT relinquishing this market."
I think for a while some didn't believe how serious Google was about the whole selling cloud services thing. Now they do.
Regarding Google and Amazon's disparate profitability:
1) Cash is arguably more important than profitability from an investment financing perspective. Amazon and Google are in the same ball park regarding free cash flow (>1B, <10B) and while Google has an enormous cash pile Amazon certainly has the ability to spend several tens of billions investing in their clouds in the next few years, as does Google.
2) Secondly, for the sake of their stock price, Google would be very hesitant to have their high margin search/ads business become dominated by a low-margin commodity cloud business. Their scope to use all of the free cash generated from search/ads for their cloud business is limited. Stock analysts don't like declining profitability whereas Amazon already has basically zero profits and is basically priced at a Price/Sales ratio like a large retailer.
http://seekingalpha.com/article/2743335-understanding-how-am... and http://www.fool.com/investing/general/2015/02/04/amazon-just... discuss the mechanics.
They're sort of taking on a bunch of debt (or things that look like debt) and downplaying it with clever accounting. In the current market environment I don't know that it's really a problem (they just borrowed $6B via a bond issue, for example). Just an interesting reality for a company competing with two cash-rich rivals in a hugely capital-intensive space.
For example, how is EC2 much better than GCE, if it is. (I realize you didn't imply that it was; this comparison just seems the most interesting to me)
AWS have ~40 instances types and and ~40 services they offer, so this is way ahead of other cloud providers, so it is a much richer ecosystem.
When I see a brand name mentioned randomly in an arricle, I wonder who paid for it. Why would Bloomberg mention instacart specifically? Is it common practice for companies to pay for these subtle pseudo-mentions in articles?
#1 only comes from a lot of ongoing PR outreach, coffee chats, briefings, etc - which takes time and $$$$ usually via an agency working nonstop on developing relationships to keep a brand top of mind for side mentions like this and to be at the top of the rolodex when a reporter needs a quote or some help understanding an industry or technology space.
Yes, much of HN seems to think this is an insightful practice. That's all I have to say about that.
Disclaimer: DevOps for AWS and bare metal infrastructure.
To get into the same cloud market as AMZN/MSFT/GOOG you are going to need:
1. Several billion dollars of Capex to get to the same scale as the incumbents.
2. A large amount of investment to replicate all the internal software that Amazon and friends have.
3. Goodwill from startups and/or enterprise CIO types.
Getting into the airline business is comparatively easy:
1. You lease the planes from Boeing or Airbus.
2. Pilots and other staff are highly skilled but otherwise fungible.
3. Marketing through traditional consumer channels and price comparison engines is very well understood and if you're willing to eat the losses then growing share is doable.
Just like getting into the airline business, and getting into the airline business to compete with the likes of American Airlines, Lufthansa or United Airlines are two different things.
The same overly simplistic business plan could just as easily be applied to 'getting into Cloud', and while leasing a few business jets off Boeing has it's own page on boeing.com, it's still a significant expenditure.
Here's the same plan:
1. Lease a pile of servers from Dell or HP
2. Developers, Sysadmins, and other staff are highly skilled but otherwise fungible. OpenStack is rapidly maturing so you can take advantage of that.
3. Marketing, marketing, marketing. Through newer online social media channels instead and compete on price.
Airlines are famous for making only meager profits, especially considering the capex required to play.
Since Cloud is still a relatively immature industry, there's still room for niches, just like the airline industry has regional airlines. All GPU cluster anybody? Bare-metal Private Servers? Digital Ocean and Rackspace are chugging along just fine.
How about selling of Cloud services that run on top of AWS? Is that still Cloud? Dropbox famously started off just running on top of Amazon's S3, and I'd definitely call them Cloud. How about if you setup Heroku for Erlang? Scala? No billion dollars of Capex required, and I'd still definitely call Heroku a Cloud business.
Trying to compete with the likes of Amazon, Microsoft, or Google without billions of dollars is tilting against windmills, but to say 'getting into cloud' requires several billion dollars at the outset is a bit of an overstatement.
You already lost the game. The entire point of cloud computing is you don't need any of the features that Dell and HP heap onto their servers. Every port, every button, every LED costs you another couple hundred mW.
AWS is eventually going to be a $30 billion sales business, capable of generating $3+ billion in net income. It'll legitimately be a nearly $100 billion market cap company unto itself in ten years. If Amazon is smart, they'll wait until it achieves a much larger scale before spinning it off, it's obviously a huge pot of gold in the equity value.
Amazon Web Services, Inc. is a separate corporation wholly owned by Amazon.com Inc
I agree, it seems unlikely in the near future given Amazon's history, but it could happen. Maybe 10-20 years from now when Bezos has retired, a new CEO or three has been in place and the public markets aren't as forgiving of Amazon as they are now.
> The AWS figures allow investors to value what some think could be a spinoff prospect.
For example, if AWS is worth $100 billion in ten years, and Amazon as a whole is only being valued at $150 billion in the public market at the time. Separate the two might be worth $100 billion each at that point.
The argument for eBay spinning off PayPal is this same premise - that PayPal is being undervalued by a lot hiding inside eBay.
PayPal however is growing much faster than the auction platform (which is hardly growing).
The old processors like Visa and Mastercard have had amazing value increases in the last ten years. Visa is now a $160 billion company with 40% net income margins. PayPayl is a very attractive target for the big four of Visa, Mastercard, Discover, Amex. That will drive up its perceived value. One of those big four is very likely to go after a merger or acquisition of PayPal after the spin-off.
As for Microsoft, quoting directly from your link: "One person has told Business Insider that Azure revenue is at $1 billion lifetime since its debut in 2011, but is accelerating quickly."
I thought that the potential profits of cloud business were already being factored in the current evaluation, but apparently it was more profitable than the analysts (which were probably imagining the same thin margins as the retail business) expected.
Tech isn't the hard part - barriers to get a service up and running are dropping drastically.
Hard part is marketing & scaling, and as AWS expands services there are risks from pure-play vendors that do just one slice of the stack 2-3x better than Amazon does.
1. OpenStack still has plenty of bugs and issues you are going to hit.
2. You'll still need a really good colo facility.
3. High bandwidth and Multiple internet providers are expensive. If you look at Amazon they have an excellent blend of bandwidth including direct connectivity to many eyeball providers (Comcast, Charter, RCN etc.) which usually makes your data get there quickly and prevents peering bottlenecks.
4. If you need more than one facility then you just get to double your costs.
5. People costs, you'll need people that can manage your infrastructure at scale.
6. Hardware capex can severely drain companies.
7. Vendor sprawl and support. (You still may need a CDN, DNS provider, etc.)
Doing your own thing can quickly become much more expensive than Amazon or other cloud providers.
FWIW, I've been told that their investment per data center is in the billion dollar range. That capital requirement is definitely a barrier to entry.
At the moment the only realistic challengers to AWS are Microsoft and Google.
And what ever happened to Eucalyptus? https://www.eucalyptus.com/
- $1.57b in revenue in Q12015 on $22.72B total ($5B annualized for AWS)
- $265m in operating income (net loss for Amazon as a whole was $57m)
Looks like revenue was slightly lower but operating income quite a bit higher than most Wall Street estimates.
Usually, application developers are not good at server side configurations. They want something that's easy to setup and help them deploy their applications like WordPress, Magento, Drupal, Joomla etc.
Companies like Cloudways are taking away a lot of pain and giving those application developers a chance to easily setup their apps with premium support.
Don't you think that's attractive? and a business opportunity as well.