I’m a new user and thought I would leave some feedback for you.
I was surprised to see the dashboard registering a cost despite the free tier. I’ve found the docs and other forum posts explaining this but I think a link from the dashboard is required. However, as a trial, I’d like to know the cost per 100GB for cold storage. So maybe think about letting users see the cost but also linking to some docs to explain the $1.65 discount thing.
Segments! what the hell! I do get it now but I didn’t realise beforehand. I’ve uploaded about 90GB of family photos and video clips, so around 8000 objects and 9000 segments.
Billing in $. I’m in the UK. I’d prefer GBP. Please consider an alternative mechanism, and not just me but the whole EU satellite could be paying in Euros right?
Bucket names in lower case only - Not too impressed with that limitation.
Bucket passwords - I like this. Similar to Onedrive and their vault, separate from the main area.
Drag and drop files - I lost files using this method! I dropped a folder of about 1500 files 22GB. Nope, I got 8GB of data. It was awful. I deleted and used the folder upload option which appeared to work.
Is there a way to check all the files I’ve uploaded are actually 100% coverage of the files I have on my drive? That is, if I upload 1500 files, how can I be sure you have them all? I need a file count. I need a folder size. Both of these metrics I could use to compare to my locally stored files. Consider an app to do this job, to help with the one off / continuous sync with cloud and local.
Thanks for the suggestions! Storj is really not an end user product so much as a platform for developers to write apps for. Or to integrate existing applications with it. It’s not meant to be a Dropbox replacement directly.
That said, there are areas that you mentioned that could be improved and/or streamlined. We’re always making improvements. Thanks again!
Most of your concerns stem from an assumption that storj is an end user service, you are even comparing it with OneDrive.
Instead, it’s more appropriate to compare it to Azure, the storage backend OneDrive is using. Segment fee is in the same vein. It reflects actual cost involved, and may influence client software design: Storing few large objects is cheaper than the same total size of a small objects. All similar systems (google cloud storage, azure, s3, have similar cost structure). The services that are built on top of them — such as OneDrive or iCloud — absorb and aggregate these costs to provide users flat-fee service, and users don’t directly interact with underlying storage service.
In other words, Storj DCS is not an end user product, it’s a storage system that other services can be built upon; for example one may build a file sharing or collaboration platform with reliable deans and drop and other features, such as flat pricing structure.
All those other restrictions — like bucket naming, lack of hierarchical “directory structure” — come from compatibility requirements with existing protocols, such as S3.
The rudimentary web interface storj provides for seeing some metadata, as do all other storage providers, is a courtesy and convenience, intended for occasional management, not daily primary use.
There is also no bucket-specific password, ( this part of Ui is misleading and is being changed to reflect it), a bucket can contain objects encrypted with different keys.
And lastly, I would not use drag and drop for uploading a nontrivial amount of objects to anywhere, including storj, especially since you already know it’s unreliable — use uplink utility or any number of third party tools like Transmit or Cyberduck. There is no reason to invoke the whole web browser stack when a couple of api calls will suffice.
Thank you both for replying. I’m shocked to hear storj is not meant for end users. But then again, something about it didn’t feel quite right. Not sure what exactly.
I had heard of storj something like 6 to 10 years ago but have never had an account until now. I think back then it was touted as some kind of crypto storage system or maybe that was just me reading about the storj tokens and making bad assumptions.
Does this mean there are active front end platforms out there? I assume there are since I have read storj is storing significant quantities of data. So I suppose I should use one of those.
My use case is effectively remote cold storage, something like Amazon glacier, but priced in a simple user way and not all that crazy techno stuff glacier is priced in. So cheap is important. Data access not important. I’m happy to wait days, weeks for access so long as it’s cheap. But my data also needs to be reliably stored such that I can trust to get it back and not get ripped off. Do you have any good suggestions?
Do you mean long term backup/archival? If so, you seem to be already using the best fitting solution, and maybe further consider transitioning to Amazon Glacier Deep Archive.
Storj is hot storage. It offers tremendous benefits that are significant for specific use cases, tailored towards fast (low latency, high bandwidth) access to decentralized data from anywhere in the world. This facilitates CDN-like behavior, and very high durability due to decentralization and redundancy. These features are not essential for long term backup use case, and while today Storj is one of the cheapest hot storage providers, it’s still hot storage, and therefore, you are still paying for all those awesome features you don’t benefit from (except, maybe, decentralization).
Regarding the cost structure—Amazon’s cost structure is fairly simple, and designed to encourage intended use of each storage tier. For Glacier Deep Archive—Ingress is free, Storage is cheap—$1/TB/Month for deep archive tier (with 180 days min retention, to discourage hot usage, and encourage long term storage, exactly as intended). You can restore 100G/month for free. Restoring more than that—can be expensive, depending on how fast you want your data back—from expedited (very expensive) to bulk (not so much). You rarely need all your 10TB of data all at once, so the cost can be spread over long time. It’s like an insurance policy.
Storj has similarly designed pricing features that intend to drive behavior. For example, segment fees discourage storage of many small files.
Ultimately, you want to optimize cost of the whole solution, incorporating the cost of restore, and not fixate on just a part of it, let alone least probable one. Since you pay for storage all the time with 100% probability, and restore very infrequently, ideally—never, your total cost is 100% * [storage cost per month] * [number of months] + [probability of total loss] * [cost of restore]. Since probability of total loss is close to zero—cost of restore is irrelevant.
For the end user platforms—for backup I personally recommend Arq7. It supports both Storj (via S3) and Amazon (including Glacier Deep Archive tier), among other things, so you can try and see what are you comfortable with. They also provide (somewhat outdated) worksheet on how to calculate cost of restore: AWS Glacier Pricing - How to Calculate the Real Cost | Arq Backup Since that was published Amazon S3 cost decreased somewhat (significantly?), but this is a good ballpark.
That said, many people use Storj as a backend storage for a variety of backup use cases, and some software (Notably, Duplicacy) supports storj natively (without relaying on a S3 gateway). And if you suspect that probability of data loss and need for a full restore is high enough – maybe using Storj as a backend is the best choice for you.