14 comments

[ 2.7 ms ] story [ 40.6 ms ] thread
This doesn't really tell me all that much - what features does the author use on Heroku? What features were available on Ninefold? What had to be changed before deploying to Ninefold? How does the hardware/environment differ?
Yeah, and how is free or $34.50 for 2 dynos vs $44 a draw?
Most importantly, why are Ninefold's monthly packages more expensive (assuming a 30-day month) than were you to use their hourly rate in their calculator?
The monthly packages include an assumed package with disk for the VMs, disk for DB, and IP addresses. The hourly rate is raw VM pricing solely.
Alright, so it's outbound traffic + app storage that you pay for. (After all, you can set the DB disk size in the calculator)

It's still a bit confusing though, (IMHO), given that the discrepancy between hourly & monthly rates grows bigger the bigger your package is. It starts off from ~0.2 USD and goes up to ~400 USD if you max everything out and there's no detailed description of what additions are contained in the monthly package.

Also, on the more detailed pricing page their "recent prices" are broken and only show if you switch to Asia + then back to the US. Furthermore, do you need to pay for the server's forced minimum storage requirements (e.g. 20 GB * 0.16 USD per ? - I'm assuming monthly here) or do you only have to pay for your additional requirements?

EDIT: also, the suggested packages on the more detailed overview page list a 512mb RAM / 1 vCPU instance which is no longer available at all - (isn't that what heroku offers for free)?

Maybe you need 3 dynos to match the performance of Ninefold's starter tier? I wouldn't have thought so though... 2 dynos will be plenty responsive and won't need to be kept alive with New Relic/Pingdom etc either.
I agree---this was one of the most content-free articles I've read that compares products.

Any time I see an article pushing a relatively unknown company over a well-established one, there's the suspicion that it's just part of a marketing campaign. And in this case, there are no benchmarks or anything substantial to back this up, even though the author tries to create the appearence of a subjective review by assigning category "winners".

"Over $100 million invested in our own hardware means big scale, better performance and a lower price per delighted user. "

This coderfactory article is actually a good advertisement for Ninefold in my opinion. I hope they charged them plenty, considering the fact that Ninefold is apparently rich. And if I were a Rails dev, I would probably check it out, if I didn't know about Digital Ocean or bash/Ubuntu stuff.

But $100 freaking million in hardware? ??

Disclaimer: I'm a Ninefold employee.

I can see how that claim could sound pretty ludicrous. A startup coming from nothing isn't going to get $100 Million to spend on hardware.

We're backed by a publicly listed Australian company: Macquarie Telecom. We've built upon their investment in datacenters and fiber connections between them as well as racks, servers, switches, routers, power etc.

We thought it was important to call out that we're not a fly-by-night operation, hosted in an office cupboard. We (and our backers) have ploughed serious capital and we're serious about what we're doing. I understand that we may not have explained very well how that investment works.

All that said, we have the advantage that most platform providers don't: capital invested in hardware which means that we don't have to rely on another provider (such as AWS) for compute resources.

OK, so I think you should say that on the site. How much of that $100 million investment can Ninefold actually access today in order to spin up servers? The way you phrased it is not truthful, and makes it sound like you built a bunch of mainly unused custom datacenters full of racks and are just waiting with your fingers crossed for the hoardes of Heroku defectors to come and load up the servers.
i've tried heroku a few times, always found it kinda painful to learn and felt helpless when things go wrong..

not sure if its because i'm used to building own servers (which might be inefficient)

maybe heroku is better for people with less sys admin skills? or it might just be way better now (i haven't tried in over a year).

now i use rackspace/ec2/linode for rails.

This is the first time I've heard of Ninefold. I checked out the service just now.

The first thing I looked for was support for multiple environments (e.g., staging and production). I searched ninefold.com for "staging" using Google, and there weren't any relevant results:

https://www.google.com/search?q=staging+site:ninefold.com

Here's an example of the type of information I was looking for:

https://devcenter.heroku.com/articles/multiple-environments

The second thing I looked for was support for zero-downtime deployments. I searched ninefold.com for "zero downtime deployments" using Google, and there weren't any relevant results:

https://www.google.com/search?q=zero+downtime+deployments+si...

Here's an example of the type of information I was looking for:

https://devcenter.heroku.com/articles/labs-preboot

After not finding the above two features, I stopped looking for other features that interest me.

Hey subsection1h, I'm Ryan from Ninefold. We're working on getting our docs in place around multiple environments. Our setup is different from Heroku and they do a great job of describing why you'd want more than one environment, so I think our article will be much different, but I'd like to make sure it's informative.

Is there any specific information you'd like to see in there?

Also, we use Phusion Passenger to manage our deploys and we do implement pre-start: http://www.modrails.com/documentation/Users%20guide%20Apache...

We're working on a new feature to further reduce the chance of diminished performance during deploys. There will be information about it in our docs and on our blog as soon as we release it.

Thanks, Ryan Clark Lead Rails Support Engineer | Ninefold.com