Amazon AWS Free Usage Tier
aws.amazon.com
aws.amazon.com
I talk to people at Amazon frequently enough to know that they don't take existing customers for granted... but doing things like this certainly doesn't help with their image in that respect.
Guess I'll be opening a new account too.
"I’ve researched your account and have confirmed your account is eligible for the AWS Free Usage Tier. Though the promotion does not actually start until November 1st accounts created from 10.21.2010 or later are eligible."
Did they tell you something different or just update the page with this?
EDIT: I pulled the wrong quote. Either way, sucks that Amazon would alienate existing customers.
These free tiers are only available to new AWS customers and are available for 12 months following your AWS sign-up date. When your free usage expires or if your application use exceeds the free usage tiers, you simply pay standard, pay-as-you-go service rates (see each service page for full pricing details). Restrictions apply; see offer terms for more details.
Cancellation is done over email, and I haven't gotten anything back yet. I used another email address to make a new account.
It is not really alienation. They are investing money in converting people to their platform. Offering this to existing users makes no sense from that perspective. It is pretty standard across many industries to offer exclusive deals to new customers.
But then again, I'm HN's resident marketing ignoramus, so what do I know? ;-)
That's why it's a mistake. Existing users (the wrong people) hear the message much louder than potential users, and possibly in greater numbers.
These expectations are analogous to you being a paying member on some platform, and they announce a "30 days free" promotion for new members, and you demanding that you should get to participate in that promotion. Shenanigans.
And just because AWS is pay as you go doesn't make this any different.
It's a way of showing appreciation and that they don't take you for granted. Period. (sorry, I had to)
If you aren't negotiating better terms for recurring services, then you are giving money away. Promotions are just a nice bit of leverage in negotiations.
In comparison if you ask your internet provider to bump you up from 1mbps to 2mbps (or to lower the price from 50 bucks to 40) because they have a promotion for new users, and they give it to you, providing you sign a contract for another 12 months, is not equivalent to telling the internet provider that because they are offering a free year of service (up to lets say 5 gigs of downstream bandwidth) to new customers you should do to and without a contract specifying that you're staying with the company afterwards.
I think the most effective type of marketing will always be the personal recommendation from a current user.
You will become ineligible for the Offer if, during any 3 month period, you do not [..] (b) incur fees for use of the service that exceeds the free usage amount provided under the Offer
So it is free as long as you exceed the free quotas. Hmm.
Micro instances are only available using AMI's using EBS. A Lucid Lynx AMI alone is already 15 GiB, and the free offer only provides 10 GB. Essentially, it's like a temporary price reduction per month.
I think this will also open the gates to developers outside the US -- $50/month for hosting might not seem too much to an American, but in other countries it can be a real barrier.
In GAE, you have the privilege of bypassing the administrative stuff - creating/deleting/configuring instances, choosing software, amongst other things. Quite useful for some app categories, like web-apps/web-services for verticals, where you want to concentrate on the business logic and the rest doesn't matter as long as the app works.
EC2 is great too, and attracts developers who require more freedom in terms of software architecture, and choice of components, but comes with a the non-trivial administrative costs. (I do imagine EC2 consultants being in-demand, sooner or later.)
I think both models have potential to succeed, and should be able to nicely co-exist. This free-to-jumpstart approach is perhaps targeted just to stir up AWS demand.
750 hours ELB = $18.75
15GB ELB data processing = $0.12
10GB EBS, 1 month = $1.00
1 million EBS I/Os = $0.10
1GB snapshot = ? (varies a lot, but less than a dollar)
5GB S3, 2000 PUT, 20000 GET = $0.79
30GB transfer ~= $4.20
The remaining items are always free.
Total: $39.96
(Marketing value: priceless)
FYI, you can calculate any bill here: http://calculator.s3.amazonaws.com/calc5.html
... except in the future, when it will be even easier :~)
Now, I tried 'free' before and I just about lost my shirt. I got a bunch of jerks signing up with fake names and then running up my bandwidth bills with torrents, then at the end of the month, nobody, ah, "converted" so I am not going to respond with 'free'
One of my responses could be to push my larger domains (which are quite a bit cheaper than the competition) or continue focusing on the international market. (I haven't read the fine print of the amazon offer, but I'll eat my hat if they give free accounts to, say, Brazilians.)
So yeah; my current attitude is "wait and see" - as of today I'm still selling out servers as fast as I can put them up... if that changes, I'll change something.
As for slicehost/linode, my personal guess is that they will both tout their substantial customer support advantages over amazon.
You really have everyone beat on price, by a mile. Too bad, there aren't enough cheap, unix-gurus to make use of it :-) People these days want an ajax console with drop-down menus for shell commands.
I have a month-old AWS account with about $0.30 in lifetime charges on it. I never quite felt like I knew how much of a bill I was racking up when it was turned on.
God! I love companies with IRC chans.
the thing of it is, really, prices for this sort of thing should be falling with Moore's law, right?
Also, there seems to be enough cheap UNIX people to keep me busy... I've been selling out hardware about as fast as I can put it up. at this point the bottleneck is pretty solidly me.
Of course, I'm charging prices that would be pretty high margin if I was 'at scale' and, well, I'm not even close, at the moment, so I don't have as much wiggle room as I would have otherwise, but rest assured, it's profitable, at least when I don't screw it up. I've set a (I think realistic) goal to double in the next 6 months, which
So, assuming the market is free (which is mostly is) and the barriers to entry are low (and they are very low.) assuming there are customers who care about price (and there are) you should see prices falling towards the cost to provide the service.
As far as I am concerned, I would just like to know you're living comfortably, that's all. If you're happy, so am I.
I'm quite new to VPS. I recently deployed a couple of project for customer in Linode. I am very happy with them, but at the same time, I have some hobby projects of my own that I dont want to host in a 20USD/month machine.
How are your VPSs compare to Linode?
I also give you fewer VCPU on the lower-end instances. they give you 4, I give you 1. the idea here is to optimize for the worst case (if there is contention, and you have 4 vcpus, that's more context switching and more cache flushes. on the other hand, cpu is usually mostly idle, and in that case more vcpu is more performance. My large instances have more vcpu.) I'm experimenting with my vcpu policy and may bring that more in line with linode. But really, I regard this as relatively unimportant because VPSs are rarely CPU bound.
Our disk quality is likely about the same (I use 7500rpm "enterprise" sata, which is about as good as sata gets without getting into the expensive 10K stuff, and if linode was using 10K disk, they'd be advertising the fact.) My ratio of disk to ram is 4 disks in raid 1+0 for every 32GiB ram/8 cores I'm not sure what Linode's ratio is. But Linode has better mechanisms for limiting heavy disk users than I do. (heavy in terms of performance rather than in terms of space... limiting by space is trivial.) My understanding is that they keep iop counters or similar, and first warn, then limit you if you use a lot more iops than most people. I don't do this, and sometimes overall disk performance suffers because of it. But I do plan to emulate Linode in that regard at some point.
And, of course, my most noticeable deficiency is that I do manual provisioning, and Linode has a nicely automated setup. but I'm (very slowly) catching up.
Now, I'm about half the price, too, but you already knew that.
My only complaint so far has been that I can't seem to figure out where I reset my key (I lost the original to a dead computer), while I still have root I don't dare reboot it because if it doesn't come up I can't fix it.
I'm a customer of yours because I know what I need, and with prgmr I pay for what I need and nothing more. It's great.
For me, the AWS free account is to more about letting developers play around with the unique features(/challenges) of AWS before needing to actually deploy a real product. There are a lot of unique things to consider with AWS that you don't have to deal with on regular VPSs.
This isn't about creating a free tier of hosting, and if Linode did offer something like this I'd be concerned about the potential impact on the hypervisor machine load and additional support impact.
* = after spending a minimum of $5,469 on a subscription
Microsoft also offers a free promotion to all new subscribers (click "view details" link under "Buy Now" button): http://www.microsoft.com/windowsazure/offers/
From the site:
Included each month at no charge:
* 25hrs small compute instance
* 500MB storage
* 10,000 storage transactions
* SQL Azure 1GB Web Edition database (3months only)
Any monthly usage in excess of the above amounts will be charged at the standard rates. This introductory special will end on March 31, 2011 and all usage will then be charged at the standard rates.
Either way, it's a pretty big platform to be locked into. Kinda like saying your new laptop locks you into a closed platform.
Is it as easy to move from Azure to a co-located windows box as it is to move from ec2 to a co-located Linux box? (e.g. install the image on the new box and replace the kernel?)
That was in cloud vs. opensource... but I hereby humbly propose 8ren's amazingly wonderful Open Service concept:
The cloud is the future, not only for the backend, but also components within it - including middleware. Theoretically, arbitrary functionality can be a service - literally, any "function" (in the math sense - or stateful). That is, anything that was previously open source.
It might even get to the point where the economics make sense to use an "open service" instead of hosting it yourself, analogous to conventional open source. This would require hosting that is: (1) practically free; (2) very high performance; (3) very high availability; (4) security/privacy mechanisms. We may be closer than you'd think - eg. if EC2 hosted "Open Services" for free, which would naturally be faster for EC2 paid services, making those paid services more attractive. Then the competition needs to match it... Google might do it too (already does, within its app engine).
To qualify for this amazing free hosted Open Service? Be open source.
http://blogrium.wordpress.com/2009/10/29/public-open-source-...
I think my suggested payment model makes Open Services easier to adopt, because it has no admin overhead. The API is there, you call it, done. It is paid for by amazon/google, because it benefits their paying customers - who indirectly pay for it. Note that they would be paying for its hosting anyway, if they deployed that open source themselves.
- There would be admin savings from having just one deployment (more efficient for everyone).
- And it would be just as fast, if deployed in the same cloud (theoretically - I'm assuming a coarse-grained IPC world, where apps are built by assembling services anyway, which is where we're headed. I'm thinking of it like static method calls in Java, that so get JIT inlined/stitched into one big lump at runtime, even though they are in separate modules (class files)).
- Plus, these free Open Services act as an advertisement and enticement to non-EC2 users (or free users) - and once you've started using EC2-hosted Open Services, the easiest step up is of course to move to becoming a paid EC2 yourself. There's a bit of lock-in. So it is in Amazon's interest, too. So it's economically a little bit like advertiser-subsidized TV. And technologically, it's like Google's many free API's (eg. map API), but the case is more compelling for the supplier, because it costs the same to provide, but more directly pulls in paying customers.
- I agree there would sometimes be issues with how much processing power it gets (Dos attacks, abuse in general), but this is a complex issue, and I think you'd need to actually run it for a time to discover the real issues and experiment with solutions. But the vast majority of open source projects are not used by anybody - these aren't using any resources (apart from memory, which is close to 0 cost), and so their cost is negligible.
- once one hoster provides it, it gives them a competitive advantage, and other hosters need to match it to stay in the game.
- over time, with Moore's law, the hoster's cost of providing these Open Service approaches zero.
I think it's inevitable for Open Services to begin and spread (or something functionally similar).
Or is it an entirely different animal?
To supplement this, they're also offering Elastic Block, which is an on-demand, resizable static partition. This is your long-term disk, but it's sold seperately, and has different usage / pricing metrics.
Generally speaking, you allocate your storage, then build your EC2 image, then create a mount, set up your application / server configuration / etc, then take a snapshot. That snapshot is the image you'll boot in the future.
Is it feasible at all to (ab-)use this free Micro instance as a real development machine for running little experiments? Assuming I keep one instance running at all times, without adding any new ones, it seems like I could just ignore the fluid snapshot system and treat it like a bog-standard VPS.
You can still use them as a standard VPS, of course, if you just take a little care and mount a static file system up front, then use that for all dynamic content/database store/uploads/etc.
The only real difference is that your persistent file store is completely separate from the instance, which threw me for a big loop when they were new, but has become old hat since then.
I don't know what the reliability is nowadays, but there used to be (when it was very new) complaints that the instances would crash or come offline without notice, and that as with all things, it would happen at very inconvenient times. With care, and forethought, it's preventable, of course; but what happened then was people trying to do exactly what you suggest, only to be thwarted by the unreliability of the system.
Of course, the beauty of it, if planned for, is that it doesn't NEED to be reliable. If you use the Elastic Load Balancer, you can spool up pre-configured instances at the drop of a hat that service demand, and if one crashes, it's easy enough to replace with a brand new instance. The instances can all share a persistent file store, they can all be dynamically registered with the load balancer. They can all function on their own with relative autonomy (which is how businesses like CloudKick make their money.)
If nothing else, I'd advise you take advantage of this deal and play around with it. It's a neat little system, and their elastic IP addresses are something of a thought exercise at the beginning, and nothing short of brilliant after you've figured it out. You'll learn something, and that's always worth something.
http://aws.amazon.com/about-aws/whats-new/2009/12/03/amazon-...
A few observations: - I wonder if you could create a new account for each of your projects (er.. startups) or if Amazon would frown on that. - They seem to be extremely generous with this offer, except for the Amazon S3 limits. 20,000 GETs? That's nothing for the kind of resources that S3 is best at hosting.
How many credit cards do you have? I suspect that if they see you using the same credit card on many accounts they'll "suggest" that you should be using consolidated billing.
"You also will be charged AWS’s standard rates for any use that exceeds the free usage amount provided under the Offer."
"You will become ineligible for the Offer if, ... (b) incur fees for use of the service that exceeds the free usage amount provided under the Offer. ..."
What is happening now in both available (and free!) cloud computing resources and the entire lean startup mentality is so far beyond my comprehension I don't even know how to compare the two startup worlds of 1995 and 2010.
Lordy!
As an aside, I realize how crazy it is to say that a year lease on a vps for free is not a great deal. This is an amazing, magical time for web developers/programmers.
That's not how I read: "You also will be charged AWS’s standard rates for any use that exceeds the free usage amount provided under the Offer."
I agree, it would be nice if amazon could allow you to put some hard limits on all of the AWS services. That's why I haven't used AWS yet - I'm afraid of that 10K bill if I should 'accidentally' hit the front page of Digg/Reddit/HN.
If you can't be sure that $20/month makes sense then you've wasted a lot of time building an app. If you don't have the $20/month then surely anyone capable of developing an app can work for a few hours on a project for someone and make more than enough to cover a year's hosting.
If you just want a shell account to dick about with then Linode isn't for you and I don't think it's unfair to suggest that us folks running businesses off Linode instances don't want your eggdrops and IRC servers on there either.
1) This move by Amazon is good because it will probably force Rackspace, Linode, et cetera to reevaluate their plans and make better offerings. This doesn't only affect the $20 plan, it affects all the plans. Hell with my requirements I'd be spending $1360 a month in Linode, in which case wouldn't you like paying a Benjamin or two less if you where in a similar position?
2) Anyone that doesn't have any infrastructure in place yet would be a fool not to take advantage of this promotion to help bootstrap their app. Hell if it doesn't work out you saved the equivalent $360 the experiment would have costed you in Linode.
3) That sadly I can't use this promotion because I already have an infrastructure in place consisting of 17 instances and other Amazon services.
Where did I suggest anything about eggdrops and IRC servers? So are you implying Linode is not for me because I need 17 servers? Did you even read the comment before trying to refute something I haven't said?
Be it $400, $100 or $50, AWS could (in my opinion should) go that extra mile and offer some option without requiring credit cards if it truly wants to take on GAE in this regard.
Only after much clamoring by people who actually wanted to pay Google money for more usage did they offer a way to pay for higher quotas, almost a year after the initial rollout.
I'll probably get an Android phone sometime soon enough, though. But I tell myself that the longer I wait, the more advanced it will be!