Re: Reliable Cloud host ?

[David Miller <dmiller () tiggee com>]

On 2/27/2012 10:25 AM, Jason Gurtz wrote:
> > [...]  For DNS,
> > EasyDNS (https://web.easydns.com/DNS_hosting.php) are rather good and
> > not too expensive, and you can get a 100% up-time guarantee if you
> > want.  A review of them regarding availability is at
> > http://www.theregister.co.uk/2012/01/31/why_i_use_easydns/
> I have been a very satisfied EasyDNS customer for about a decade and
> concur with the article. Nothing is perfect, but the rapid response and
> support I've received have always been top-notch.

I have been a satisfied DNS Made Easy customer for many years.

Note: I am also an employee of DNS Made Easy.  I was a customer for
years before I became an employee.


> > Do let us know who you end up picking and how it goes.
> Indeed. "Cloud" outside of references to mists and objects in the sky is a
> completely meaningless term for operators. In fact, it has made it harder
> to differentiate between services (which I'm sure is the point).
>
> As an operator (knowing how things can be subject to accelerated roll-out
> when $business feels they are missing out), I wonder if a lot of these
> "cloud" service bumps-in-the-road aren't just a symptom of not being fully
> baked in.

It depends on what you mean by "bumps-in-the-road"...

If you mean issues experienced by customers of cloud service providers,
then the most common issues are a symptom of not implementing redundancy
(anticipating failure) in their usage of the platform.  There are a
whole lot of folks who believe that they can buy an instance from Vendor
=~ /.*cloud.*/ and all of their DR worries will magically be "taken care
of" by the platform.  That isn't the case.

Amazon is usually pretty good at providing RFOs after issues.  All of
their RFOs (that I have seen) include pointers to all of the Amazon
redundancy configuration documents that customers who did experience an
issue regarding the RFO did not follow (which caused them to experience
an outage due to a platform issue).

DR in using cloud services is the same as DR has always been - look at
all potential failures and then implement redundancy where the
cost/benefit works out in favor of the redundancy.  Document, test,
rinse, lather, repeat.

Rightscale and other services like it provide tools to help.

-DMM