Sync with Amazon S3

Hi Devonthink

I would like to request that sync can happen with Amazon S3.

This is because Amazon S3 storage costs are quite low so its a very good service to use for file storage.

In order to sync DT now, I have to use a FTP or Webdav server. I have about 10Gb in various databases. I have a server setup in my office to provide Webdav but the internet speeds are low. So ideally I would like a server in the cloud somewhere but renting a VPS to provide a FTP/Webdav would be costly just for this.

So Amazon S3 is the ideal solution to have a fast and cloud based file system to store Devonthink syncs.


+1 for @amityweb’s suggestion

… however, it’s also useful to know the cost involved with S3. For example, this posting from Gabe Weatherhead for the long-term storage (Glacier) option on S3. Storage is cheap, retrieval is shocking.

And… do you really need a cloud Sync? If you are routinely at work then at home, and there’s no security policy being breached, why not grab a big thumb drive or a portable USB powered drive and just create a local syncStore on it? It’s going to be much faster, more reliable, completely in your control (ie. private and secure), and takes so little effort.

@korm, I may be wrong but I thought Glacier was lower cost storage but costs more to get it. So a frequent GET, POST etc. would put Glacier more expensive that S3. So I am just referring to S3.

Sorry, this is going backwards than what I have now. I am not carrying a drive with me everywhere, and plugging it in etc. We are in the cloud era, this is old school and I work much better using cloud and sync technology. And if I forgot my drive I’m stuffed. I HATE having to go in my bag to get my bank security device everytime, which I have lost and forgot a few times, it puts me off wanting to login to my bank. No, online sync is the way to go. I dont think Dropbox would be as popular if it was called “Dropdrive” which is a 5Gb USB drive you carry around with you :slight_smile:

That’s my understanding of its pricing.

Thought so, which is why korm’s Glacier-centric reply surprised me. :slight_smile:

Your original post implied that to me, which is why BLUEFROG’s suggestion surprised me. :slight_smile:

I’ve briefly tested A3 storage with the Arq client but have postponed long-term usage at least until its (and maybe Glacier’s) cost/value ratio is worthwhile enough for me.

Oh brave new world! I share some of your appreciation of the cloud. I like the ability to share data between my Macs and iOS devices, examples like Calendar, iWork and PhotoStream. Dropbox provides a convenient way of putting up files for access by those with whom I wish to share them.

But when I think about data security, recovery speed and recovery cost, I really don’t want to trust my data to the cloud.

Over the years I’ve lost data stored on the cloud because the service went out of business or was discontinued (I can count at least 7 times that happened; fortunately I had local backups). Even financially sound big players have had service problems that resulted in data loss.

I’ve got terabytes of data. Not all of that is critically important to me, but it’s easy and relatively inexpensive to me to keep it all backed up on a RAID storage device.

I’ve got over a hundred gigabytes of data that I do consider critically important, and I want to retain access to it at all times, and to be able to quickly recover it if Bad Things happen. Cloud storage is too expensive, too slow and too unreliable to satisfy me. That’s why my most important data is stored on portable drives and a set of copies is kept (and periodically updated) on drives in a bank safety deposit box. If all my computer equipment were stolen or lost, e.g., in a fire, I can get a new computer and be back up quickly and surely – more cheaply and quickly than if I had depended on cloud storage.

It all comes down to one’s feelings about responsibility for important data. I want to accept that responsibility myself, rather than trusting others. Perhaps I’m a control freak. :slight_smile:

I heavily prefer having that responsibility with my important data, too. I still don’t trust “the cloud” and its fickleness that can be more trouble to deal with than “cloudless” alternatives.

Is there any doubt? :slight_smile: At least none that I can be one. :slight_smile:

I understand the issue with Cloud security, but so do cloud providers and they can o a better job than me at securing it! But I think the point is I should be able to decide that for myself, and it should not be a reason for DT to not add it. Others should not be making that decision for me.

So I do choose to put it in the cloud, and to be honest I think Amazons security would be better than my USB I may lose (I have lost my bank security device a couple of times).

Also, cloud speeds are faster than my current setup as I have to connect from slow broadband at home, to slow broadband at work, whereas to Amazon it would be quicker and more reliable.

@sjk @amityweb, it might not appear that way, but normally I can actually read without being read to. 8) I actually comprehended the original post and its S3 reference. Probably that’s why I seconded it, regardless of the angst over clouds. (Would be nice to get a factual technical response, or Y/N to the OP’s request, btw. I’m curious to know why or why not S3 would or wouldn’t be considered by DEVONtech.)

BTW, my mention of Glacier (which is a sister product of S3) is meant to be cautionary – in the vein of be careful what one asks for. :unamused:

Jim’s and my remarks don’t imply that S3 will or will not be compatible with DEVONthink Sync.

Each of the cloud services has it’s own little quirks, requiring testing and (often) development. So long as that’s the case, the developers will tend to focus on services that are most often requested by users. Dropbox was the initial winner because of its popularity.

In the world of hardware storage (HDs, SSDs, thumb drives) Mac users usually don’t have to worry about formats other than HFS Plus or Fat. Those standards are pretty rock solid. RAID setups are likewise predictable. If a direct connect storage system works under OS X, it will work with DEVONthink.

Introduce networking, and standards may get looser. For example, not all NAS setups are equal; some behave transparently and others are quirky.

Currently, Sync development for the cloud is rather analogous to the situation if it had to be tested and modified to work with every brand and model of HD because of differences. I’m sure Nathan dreams of a uniform cloud, but probably has waking nightmares with the current reality. :slight_smile:

Well said, Bill. Though I am not a personal fan of the cloud (and many of you would reconsider your love if you had to deal with it from this side of the table), the technical difficulties are a fact of life. There is no ISO Sync standard. There is no agreement on the best method for data transactions like this. Each service has its own group developing their own methods and APIs so it’s not like knowing Spanish will help you to intuit some Italian or French or Latin. It’s like learning Spanish and finding out the people you want to talk to speak Rwanda-Rundi or Mandarin, etc. And each new service that crops up (and often disappears shortly after) has “the best way to sync” - adding new languages to the mix.

On the other hand, direct peer to peer connections on a LAN or via drives (portable or static) are a time-tested, controllable, and solid way to handle these transfers.

Just because a technology is available doesn’t mean it’s necessary or the best option to use.

PS: Most of the Sync problems we’ve seen have been due to network (ie. Internet) and cloud service issues, not our technology (not to say that our Sync is perfect yet). These are just the facts based on the Support Tickets we have.

Amazon S3 is absolutely critical for us.
As a growing law firm we rely on Amazon for reliable scalable and secure solutions.

We have been testing DEVONthink for nearly a month now with using amazon s3 as a sync location and it’s been working great. Although we can hack the workarounds to make it work I really would like to see this feature as part of DT releases so we don’t have to have third-party applications running to make the storage connection work.
We do not want to run webdav servers and really for a professional environment amazon is ideal due to it’s reliability and scalability.


I’m also into Amazon S3 and also successfully implemented the Local Sync Store option for my purposes. But you should take into account that you can’t use this scenario to sync both DEVONThink desktop and DEVONThink To Go (mobile app), at least currently.

Welcome @pavelbrilliant

Yes, that is possible to use a local sync store in this situation, but also correct, not with DEVONthink To Go.

Adding support for specific services isn’t trivial and also leads to extended support when things go wrong ith that service (aka people always assume an issue is ours, not the service’s). The request is noted however.