How many products is 'enough' for Amazon, before they begin to consolidate?
I can't be alone in thinking the vast range is off-putting, not to mention the more range there is the more AWS-specific it is, making it simultaneously harder and more important to figure out the right choice...
Can't you just pick the product that does what you want? If you need a WAF, use their WAF, if you just need a load balancer, then use an ELB. If you need a data warehouse, use Redshift, if you just need to make simple queries against data stored in S3, use Athena.
Having a wide range of products at a variety of price points and capabilities sounds better than a "one size fits most" approach.
There are reasons to be alarmed at the "throw stuff at the wall and see what sticks" approach.
Like perhaps you decide to use a product that they're not particularly interested in supporting very well, and a year later they decide to shutter the service which performs a crucial role in your environment...
I am not suggesting that Amazon is actually going to do this, but it's certainly more likely than, say, EC2 going away.
AWS has sunset features before (SimpleDB for example) but only for new customers. Old customers that use the EOL features 1. get a migration path 2. keep the features in their account.
Elasticache is still there, and it is still getting updates and new features: Redis got bumped to 3.2 a few months ago, and they also added managed sharding support.
Yes, this is the way regions develop. The grandpappy with everything is us-east-1 (well, historically anyway, which is why it also gets all the outages...), and the services slowly migrate out to other regions from there. You'll see new regions start to pile on services as time goes on.
Often they'll just pop into existence quietly. Our Sydney setup only had two AZs, and a few months ago, I noticed a third one. No idea when that came to life, but it would have been useful a year before :)
Hrm, fair enough. I thought it was us-east-1, though it seems that recently they're going in a few locations. But definitely, new services appear in the more mature regions first, and then gradually filter out to the others.
I'd expect there's an MVP of services AWS stands up in new regions, and then expands out from there as the underlying primitives shake out (this doesn't apply to GovCloud though).
Wow I had no idea it only takes 1 month to plan and build a a brand new datacenter!
In all seriousness, no. There's no way in hell Amazon would rush something like this just because of the election results. The risk is too high compared to the (probably very little) reward.
Oops I didn't mean to imply they built it from scratch in the last month. I wasn't clear. I meant they may have rushed the final steps a few months. Not possible?
Following the new AWS Canada region a week earlier.
An application i maintain at $DAYJOB has a drop-down selection for AWS region and i have periodically updated the list of possible values. Is it idiomatic for end-user software to manually enter the endpoint?
Any idea where the datacentres are? I know that amazon doesn't disclose it, but should be noticable to have 2-3 new huge data centres? Would guess 1 AZ in Slough and the other one in the docklands (if they have 2AZs)?
They appear to have 2x100G links at LINX to Slough Equinix, so I'd guess they have at least 1 AZ in Slough space leased from Equinix. I hope they know that Slough isn't in London.
It's nice AWS focuses more in the EU. But how does AWS align with EU data protection laws now and in near future (in the regards of being a American company operating in the EU).
I was wondering why so many announcements of non US data centers all of a sudden, in some cases very limited. My guess is that Trump winning has accelerated demand for companies to move their data off US property, and cloud providers are scrambling to meet that demand.
I don't think so. Planning, constructing a data center and setting up all the hardware within and the connectivity to the data center is a massive investment of time and resources. These newly announced centers have been in the pipeline for a long time, long before anyone had any idea about the outcome of the election.
The more likely cause was that cloud providers could see the trend towards countries instituting data residency requirements, which was clear last year.
Here is a list of all AWS regions where EC2 is supported, since I have to support new regions for my own startup (https://commando.io).
UPDATED: I also added the hourly cost (in US dollars) of a c4.large instance in each region to compare. I picked c4.large since it's a nice starter instance for "webish" workloads.
$0.10 - US East (N. Virginia) [us-east-1]
$0.10 - US East (Ohio) [us-east-2]
$0.124 - US West (N. California) [us-west-1]
$0.10 - US West (Oregon) [us-west-2]
$0.11 - Canada (Central) [ca-central-1]
$0.11 - Asia Pacific (Mumbai) [ap-south-1]
$0.114 - Asia Pacific (Seoul) [ap-northeast-2]
$0.115 - Asia Pacific (Singapore) [ap-southeast-1]
$0.13 - Asia Pacific (Sydney) [ap-southeast-2]
$0.126 - Asia Pacific (Tokyo) [ap-northeast-1]
$0.114 - EU (Frankfurt) [eu-central-1]
$0.113 - EU (Ireland) [eu-west-1]
$0.119 - EU (London) [eu-west-2]
$0.155 - South America (Sao Paulo) [sa-east-1]
How many products is 'enough' for Amazon, before they begin to consolidate?
I can't be alone in thinking the vast range is off-putting, not to mention the more range there is the more AWS-specific it is, making it simultaneously harder and more important to figure out the right choice...