Not sure how you get from 100GB to 4.5$.
Tardigrade costs 10$ per 1TB of storage. And you only pay for egress, not ingress. So if you never download your backup, your 100GB will cost you only 1$ per month.
However, I was searching for backup solutions too these last days and also came to the conclusion that there are cheaper options to back up my data. Some of those options even don’t have egress costs if you use a 2TB storage plan.
But my backups will be more like 300GB so in that size range there are a lot less cheaper options but still many cheaper options than tardigrade.
It also comes down to what features and redundancy you are looking for as well as download costs and how long it takes for the data to be available if you need it (e.g. amazon glacier).
STORJ has the advantage that your data is always accessible at fast speed and is stored more decentralized than your typical cheap storage provider.
But why download it? The download itself does cost 4.5$, so you are actually paying 5.5$ each month.
In that case you really are better off with most other solutions. Even strato hidrive only costs the same for 250GB, b2 backblaze would be 1.5$ even with downloading the whole backup once a month.
The downside of both those options however is that your backup is stored centralized in a single datacenter (afaik), so if that got hit or gets offline, your backup is gone.
I honestly have no experience with duplicati yet. Just tried it yesterday and there is an option to not verify the backups, which is e.g. needed if you use s3 glacier.
It’s always a risk to simply trust the backup solution without verifying the data. But it’s also the cheapest option.
Theoretically it is not neccessary to verify the data if you use tardigrade because it is designed to not lose or corrupt your data. But so are most systems…
It does however bother me that I wanted to support tardigrade and their “aws competitive” offer, but in fact, tardigrade is quite expensive compared to most solutions. AWS in the fastest option is more expensive, google is more expensive, Azure (i think) but most other solutions are at least cheaper in egress, many backup solutions even in storage costs.
And mainly the egress costs make a difference. 45$ per TB is just insane if you want to use it for backups. It’s great for me as an SNO but I won’t use tardigrade for my backups. With egress costs like this I wouldn’t download my backups unless I need to restore it. And if I never plan to download the backups anyway, I can use cheaper options like aws glacier to store my backups.
But many other cloud storage solutions have additional network fees (sometimes hidden) and can get even a lot more expensive than 45$ per TB. so… you have to be really careful to find a good solution for your use-case.
Tardigrade offers hot object storage. It is in fact very cheap if you compare it to similar offerings. But hot object storage may not always be what you need. Obviously glacier is a lot cheaper, but you can’t rely on that if your business continuity depends on having data available fast in case it needs recovering. I think you’ll find that other solutions with lower or no egress costs simply can’t come close to the speeds offered by Tardigrade.
absolutely. but that is not the use-case here. If you specifically look for a personal backup solution, then tardigrade is very expensive. Even using dropbox is 10$ for 2TB with no egress/ingress costs.
But these are either small personal backup use-cases or archival use-cases.
Certainly nothing you would use in a business environment. (maybe glacier for long term archival but in terms of fast hot storage tardigrade is doing good).
I honestly think that SNOs wouldn’t want to be used as backup infrastructure as well, as that means lots of disk space usage with less to nothing egress. So for backup, it’s a loss for both tardigrade user as well as SNOs.
Better to sell tardigrade as hot object storage for business solution that requires unlimited scalability as well as decentralization. CCTV recording storage, Nextcloud storage, Video Streaming platform storage, those are all good use case where everyone win.
The thing is with the current cloud storage marketplace, if a company has a lot of egress they will use something like Backblaze B2 ($5/TB/month) with Cloudflare in front to eliminate transfer fees or Wasabi ($6/TB/month) with no egress costs. Tardigrade pricing is more conducive to backup storage due to the high egress pricing. Even though AWS is the current cloud king, they are not the only benchmark that should be measure against in the long term for viability of this platform. Plus, a reduction in egress costs may result in more egress traffic as a result.
I say all of this in the hope of improving the platform as a whole, I have a decent amount of time put into this project over the years and a strong desire to see this platform become the benchmark decentralized storage platform.
@jensamberg
well you already have your data, why bother downloading it again to verify it, tardigrade continually audits data on the storagenodes, basically all you have to do is verify that all the files you want backup’s of is up to date… and if all of your files change so much so that you have to replace all the data, then i’m pretty sure you are doing the whole backup thing wrong… maybe in an antiquated way or something.
with something like zfs you can snapshot and then you just need to upload the data changes since the snapshot to update your backups.
sure it does depend on you trusting tardigrade to keep your data safe… but than again, you really should have a third offline backup somewhere anyways… thus allowing you to trust stuff like tardigrade to make your backup method easy and swift… the cool thing about tardigrade is that you can essentially have a corporate datacenter and take a backup of all the data… and then if the entire center goes offline you could essentially run the data parts of it directly from the tardigrade cloud without anyone on the user end of your product complaining about it being down…
Well if you use ownCloud/Nextcloud storage with syncing enabled then the cost of egress bandwidth will be huge, especially in case of organizations. Based on the current price it seems that Tardigrade is not the best option for syncing among team members…
Well if you use ownCloud/Nextcloud storage with syncing enabled then the cost of egress bandwidth will be huge, especially in case of organizations. Based on the current price it seems that Tardigrade is not the best option for syncing among team members…
Never said it’s gonna be cheap, but there’s a value proposition here:
in Tardigrade, we can be sure that data will be decentralized instead of hosted inside one datacenter. S3, despite their size, have had several failures that are causing data loss due to massive power outage on the DC.
Privacy, you can be sure that your data is safe from prying eyes, unlike when you host everything on GDrive where it’s going to be data mined and used against you.
Scalability.
For companies, this value proposition can outweigh the extra spendings.
One of the major problems with the current “Free” Internet is that the pricing includes the unmarked sale of the customers’ personal information.
The challenge is to identify that hidden cost to the consumer and quantify it.
Not OP, but my $0.02 on this would be to hit pricing equality with Backblaze B2.
SNOs are payed $1.5/TB/month, customers pay $10/TB/month. With a 2.7x expansion ratio, that means that for every TB stored, SNOs are payed $4.05/TB/month total. Not sure if the satellite taking a 60% cut is entirely required, but I don’t know the economics of running a satellite but as most of the bandwidth is directly customer -> SNO, not sure what overhead the satellite needs. If the economics could get down to $5-6/TB/month for customers that would bring it line with the most competitive cloud offerings from the mid-size providers.
For egress, $45/TB is a lot, especially if SNOs are being paid $20/TB for customer egress. Ookla, while only one source is a source nonetheless, shows the global average bandwidth of ~75/40 Mbps. That means that fully saturated upload for the month would allow people to upload ~13TB/month. At $20/TB, that is $260/month, more than enough to pay for internet, power, and even ~4TB of additional storage just from bandwidth alone. Based on low network repair amounts, I don’t think that repairs should be paid either, payouts that are in escrow and not payed out to a node not gracefully exiting go back to the satellite. That in and of itself should help reduce costs. As a worst case the amount held in escrow is paid out to the SNOs that took over egressing the repaired data, not a fixed amount per TB egress.
Again, just my $0.02 on pricing. I have tried to get devs and some small companies in my professional network on board but it is much cheaper for them to encrypt their data and upload to Wasabi/Backblaze.
I love the amount that Storj pays out to SNOs as a SNO myself that runs Storj infrastructure in multiple places, but I would rather take a cut in bandwidth payouts down to a couple $/TB egress and see a lot more egress. At the same time, I can understand taking a slower approach to growth with higher pricing in order to grow the SNO base and solidify the software.
For a retail Tardigrade user, who does not care at all how much node operators earn (I feel sorry for SNO guys but that is the reality for most of the Tardigrade users I guess), using Tardigrade with egress bandwidth fee is like using metered Internet line. You always have to think about the fee and be careful not to use much (avoiding Youtube/Netflix). If you forget to check the fee periodically you might be shocked by unexpected amount of fee on your Visa bill
You have to be also very familiar with the service with which you are going to use and make sure that it is really working as you expected (how the service handles the data). @VinC FWIW I have tested Tardigrade with Nextcloud already and experienced a problem, which is why I mentioned Nextcloud/ownCloud above.
Since there is no trial already, with one mistake (no matter whose mistake it is), the amount of fee can be a disaster for users. To integrate Tardigrade with other existing services you would need a thorough testing so that you would literally not go bankrupt.
Of course that is not limited to Tardigrade, but the point is unlike other non-distributed services Storj team is practically not in position to waive the fee caused by a bug or a mistake, because users owe node operators, not the team. Since Storj team needs the node operators as well, everyone expects that the fee should be paid no matter what.
I understand you pay for what you need, like scalability, decentralization, speed, etc. but it is valid as long as you are perfectly sure that you control the stuff, able to calculate the fee rationally, and are insured against fatal error. Since you have already the stable option for privacy like Cryptomator, which encrypts your data locally to upload to cloud services, I am not really sure Tardigrade is for retail users (like AWS is not for them as well).
(By the way would you please create a new category for Tardigrade users? Currently discussion about Tardigrade is under Engineer Discussions but since the general availability it should be open for users as well, not only engineers / developers.)
I’m sorry but this is just how every other object storage works too.
Sure, STORJ should definitely implement some customizable limits and alerts so you don’t get over your budget accidentally.
But apart from that, this is the same with aws, backblaze, azure, …
Every object storage has an egress free.