OpenIO: object storage and grid for apps | Hacker News Reader

OpenIO: object storage and grid for apps

openio.io

53 pointslormayna23 comments

Here's my feedback as a person being marketed to:

After trying for a while to scroll through your landing page despite the extremely annoying custom scroll thing, I have no idea what your thing actually does.

I know you have offices in France and the U.S., that you sell something involving the storage of objects, and that there is a "storage revolution" going on.

When I click "learn more", I get instructions for installing VirtualBox and Vagrant...

Even if I click further and read your PDF, the introductory paragraph tells me that you are using "the groundbreaking concept of 'Conscience'" which I have no idea what it is.

After reading a bit more, I still don't understand what your thing actually does.

It sounds kind of cool from all the cool words and stuff, but I'm never going to install it or recommend it because I can't get a concrete and tangible idea of what it is.

My suggestion: on the first page, you should have a terse description of the steps a user or company would go through and what they would gain from it. Something like... and I'm just making this up:

1. You're storing lots of objects but your hard drives are overflowing and breaking all the time.

2. You install our product, which is a server that runs on Linux computers.

3. You configure it to use a variety of storage engines, for example plugging in your S3 credentials and some other things.

4. Now you can use a REST API to store content-addressed objects, and the product takes care of datacenter distribution, backups, and stuff.

5. Benefits cascade upon you like you could never imagine, for reasons X, Y, and Z.

I noticed the annoying scroller and just closed the tab. It's just really annoying.

From the title, I don't know what they mean when they say "grid". When I think grid, I'm usually thinking about layouts.

https://en.wikipedia.org/wiki/Grid_computing

Same here. If you mess with scrolling, I'm going to close your site. Every time.

Yeah, fully agreed. I saw a lot of buzzwords, a lot of revolution, not a lot of substance on what I was looking at.

I wonder if they just bungled the translation from French to English? The marketing blurbs under "Easy", "Scalable", and "Powerful" are pretty hard to read, almost like they were output from one of those online translation systems.

Same impression. Is this comparable to e.g. http://ceph.com ?

Looks like it follows Amazon S3's design closely, rather than that of a classic distributed file system:

* Immutable named objects

* Addressable by chunk (each object has a maximum size)

* Optional per-object versioning

* HTTP API

In particular, while you could write a FUSE driver for it, it's designed for apps to use via an HTTP API, just like S3. It even implements the S3 API.

So, like Riak CS then?

Yes, though CS is called Riak S2 now.

That seems to be what they are going for but without a better site I'm not really that interested. If you looking through their PDF they talk about scaling from 3 VMs to a larger cluster.

Products like these need extremely good documentation to be worth it.

As people have pointed out, this is similar to S3, GCS, etc. and apparently they've been running it for a bit:

> The first production ready version was built in 2008 and the first massive production of a large scale email system started the year after. Since then, the solution has been used to store 10+ Petabytes, 10+ billions of objects, at 20 Gbps of bandwidth at the peak hour, with low-latency SLAs enforced 24/7.

That's not a bad scale for such a project, but an individual large customer can easily consume that much bandwidth and storage like the guys at Descartes Labs (http://googlecloudplatform.blogspot.com/2015/11/startup-spot...).

It seems like they need a lighthouse customer to really push it. Best of luck gdelaporte and team!

For those looking for a bit more detail: http://openio.io/wp-content/uploads/2015/07/OpenIO-CoreSolut....

Thanks for your feedbacks, we appreciate it. Work is ongoing for a website v2, we'll take them into account.

If you take out the scroll thing, this crowd will probably be much happier immediately!

Other than that, I just looked at Stripe.com to see how they do their landing page, and they make it pretty clear from the outset what their service is. They're doing good so they might be a good model.

I know these website UX critiques can come off as annoying, but I hope it's actionable and that you can get the word out about your product more effectively.

How is your product different than the many many competitors?

Hi hemancuso,

Maybe you will find this useful:

1/ Pure software for any mixed hardware optimized by dynamic data placement 2/ No rebalance and no perf impact on production when scaling 3/ Grid For Apps: distribute & run any apps on same nodes

If you want to go further we will be at the OpenStack Summit, feel free to visit us. We have a booth. Don’t forget to vote for our presentations: http://openio.io/events/openio-will-be-a-sponsor-at-openstac... See you there!

Is this trying to be OpenStack Swift or leverage OpenStack Swift?

They are a swift competitor with a swift compatible frontend

That's an unusual way to pronounce "data"

Yeah sounded like they got a non-english speaker to do their intro. That or someone on the autism spectrum.

I wouldn't trust my data to a vendor that can't even pronounce data correctly...

Both pronounciations are correct:

https://www.quora.com/How-do-you-pronounce-the-word-data-cor...

Its not dahta or dayta that is the issue. Its the slow irreconcilable way the speaker voices the two syllables with an emphasis on TUH. Its a soft word, not Dat Tuh with a hard tuh. Literally sounds mentally challenged and unless the team behind this product is one person who has a throne, I can't think of a reason why no one would voice opposition to the anti social pronunciation. Laughable error.

Reply on news.ycombinator.com