Hacker News new | ask | show | jobs
by dcsadmin 2970 days ago
I've built one of these several times for larger companies. They need a place to store millions of documents that integrates with their multiple existing internal systems. These documents would come in from fax, email or integration services.

A benefit this has over cloud storage is that you can store your files across cloud vendors; so it would have cloud vendor redundancy. If Amazon goes down, it will pull files from Azure, and vice versa. This is a future feature if this gets any traction.

Other things it might have over a cloud storage is you can assign tokens with different access levels, so your applications that need read only will be assigned read only tokens and your applications that need read/write can use read/write tokens.

Larger customers that aren't interested in the cloud or are cloud averse with their sensitive documents can do a local install and have it store on a share inside their own network, so the method of storage is mutable; Azure, AWS, SMB, even FTP.

You can also assign searchable metadata to each document, or upload a bunch of documents with the same metadata. So say you have a referral of some sort and it has 20 documents; you can upload all 20 and assign a meta tag as ReferralID: 123, then reference those documents in searches.

There is also a built in file conversion.

2 comments

>A benefit this has over cloud storage is that you can store your files across cloud vendors; so it would have cloud vendor redundancy. If Amazon goes down, it will pull files from Azure, and vice versa. This is a future feature if this gets any traction.

azure has secondary region built into their blob solution, which is essentially what you're describing. and considering your site has a bunch of moving parts (multi-cloud, file conversion, search, etc.), I'm not really convinced that you'll be more reliable than the cloud provider. in other words, you'll be the cause of downtime more often than s3 failing.

>Other things it might have over a cloud storage is you can assign tokens with different access levels, so your applications that need read only will be assigned read only tokens and your applications that need read/write can use read/write tokens.

aws/azure iam handles this.

>Larger customers that aren't interested in the cloud or are cloud averse with their sensitive documents can do a local install and have it store on a share inside their own network, so the method of storage is mutable; Azure, AWS, SMB, even FTP.

but there are many storage appliances (what you're essentially describing) out there, including open source ones like owncloud. what makes yours stand out?

> azure has secondary region built into their blob solution, which is essentially what you're describing

How does Azure's secondary region capability provide cloud vendor redundancy?

It doesn't. But I don't think that's an issue considering providers wide outages are exceedingly rare. Even last year's massive AWS outage was isolated to us-east-1.
A providers outage can also be named a price hike, a change in license or a security breach. In that case you might want your data out of the provider asap.
Its a nice feature but not relevant enough.

My experience is, that a company who decides to use AWS (or whatever) only decided to use one specific vendor and would need to reevaluate every other vendor the same way.

If you can't trust AWS enough, i wouldn't use AWS.

if the software is cloud agnostic, wouldn't a simple copy operation between the cloud providers suffice?

>security breach

if you want to reduce your attack surface, copying your data across multiple providers is counter-productive. instead of having to hack your specific cloud provider, the attacker only needs to hack one of many cloud providers.

In some cases a copy might not be viable, especially with price hikes. If AWS increased storage and outgoing data costs greatly then exporting all your data will most likely cause a much larger bill.

Splitting your data across multiple vendors protects against all kinds of failures from all sides (including the vendor simply closing your account for no reason)

While true, but the UI for these is pretty bad afaik.

There is value in improving the UI for less experienced devs / non technical users.

Might be worth exploring other markets?

Ease of use is certainly something that was in my mind during development. With bitbox.co, you just have to create an account and start dropping files to get going. You can add key/value metadata as you drag and drop. Of course deeper integrations will become more complex, but even then, I want to build a simple "upload/download/search" option that is only a few properties of complexity.
I don't mean to be harsh, but the number one feature you just described to differentiate your product from existing offerings... doesn't exist yet? I really don't see how your product is any better than the multitude of other storage options that already exist, and on top of that your pricing is super steep (comparatively).