Well, there are actually laws around the world. And in EU and my country within it one part of an agreement can't single handily change the terms of the agreement. The parties would have to agree about the new terms. But Google just don't care and assume they will not actually be challenged on it.
I'm curious to hear how you calculated this (i.e. what rack, drives, other hardware are you looking at and which colocation facility)? If it's accurate, I'd be interested in going this route rather than playing whack-a-mole with cloud providers. Also, did you factor in any bandwidth limitations from the colocation facility?
I got the same email a few days ago and got another one today stating they've limited my bandwidth usage for the calendar month. I tried sending an email to fairusepolicy@box.com as stated in the email, asking for them to excuse my bandwidth usage temporarily as I migrate all my data over but no response.
The free MEGA plan sucked big time (lots of throttling, errors, etc). I would not trust their paid plan to be much better.
A strategy that could perhaps still work for me is to utilize the fact that Google allows me to be over quota, as long as I pay my bill. And, that my users are billed on a daily basis.
This means that there is no need to keep my primary subscription in a writable state all times, I can get by with only adding the necessary users when I need to upload to my main drive.
So a way forward could perhaps be like this:
-
Setup domain A as a Business Plus subscription (primary storage subscription)
-
Setup domain B as a Business Plus subscription (buffer storage subscription)
Domain A will be the primary storage domain, my main drive. This drive will be in a read-only state most of the time, having usually only one user (the admin user). This also means that domain A cannot be used for google docs/google photos/e-mail/other services that need storage available at any given time.
Domain B will be the secondary, always writable, "buffer" domain.
- Setup a rclone union / UnionFS between the main drive at domain A and a shared drive at domain B. Domain B will be the writable drive.
I am now able to write my first 5TB to the rclone union - 5TB which all goes to the shared drive in domain B.
Once the shared drive nears 5TB, it is time to move the data to the main drive at domain A. As I am not over quota yet, this is straightforward. The admin user of domain A has been added to the shared drive at domain B, and simply logs in and drags the data from the shared drive to the main drive.
Now I am able to write 5TB again to the shared drive at domain B. And soon it is time to do the transfer again.
But - at this point I am no longer able to just add the data to domain A, as this drive has only one user, equalling a 5TB quota. And these 5TB are already in use by the previous transfer.
So, at this point I will need to add a temporary user to domain A, raising the quota to 10TB.
- Add temporary users to domain A, as many as needed to raise the quota to contain the added storage from domain B.
With the new user added, I am again able to move the data to the main drive at domain B. And, I am again able to write 5TB to the shared drive at domain B.
- Transfer data to domain A from domain B, and delete the newly added temporary users from domain A.
After the data is moved, I can delete the temporary user from domain A again. This means that domain A will now be in a readonly state, as it is over quota (10TB is now used, and the quota is 5TB per user).
Now I just repeat ad libitum.
As storage grows, I will of course need to add more and more temporary users when it is time to transfer from the buffer drive.
Let's say after a while I have 1PB storage added to my main drive - this means I will now need to create 200 temporary users for the duration of the transfer from the buffer drive.
It comes with a cost for me to have these 200 users active - $120 per day, to be precise.
200 users, $18/month, 1 day = $120.
The transfer itself takes only a few minutes, and the users can be deleted again afterwards.
At a point it probably becomes feasible to be able to "drag in" storage in larger quantities than the original 5TB at a time - meaning adding more users to the buffer drive subscription. All dependent on how many TBs I add to the buffer drive daily, etc.
The reason for using domain B as a storage buffer is that there is no 750GB/day limit when it comes to moving data between drives. This makes the transfer simple and fast - as apposed to using for example a local NAS as storage buffer.
Anyway, I am just speculating here. Perhaps there are hidden limits to how many users can be created/deleted per day, or other pitfalls making this not a way forward at all. We will see.
How much is "well under"?
For 320TB (raw), you would have to use at least 16 x 20TB HDDs -- which currently costs at least -- refurbished and from an unknown 'brand' -- $230 each one, so 16 * 230 = 3680 and that's just for the HDDs. Add at least $1300 for an -- also used -- server capable of taking in all 16 HDDs, and we're talking around $5K just for starters. And that's before the colocation costs, eventual "hired hands" costs for replacing failed components, cost of replacements when the originals go out of warranty (which is probably very short to non-existent with used/refurbished stuff), and so on.
And unless you're a real daredevil with your data, those 320TB raw aren't going to be 320TB usable: you'd have to dedicate some of these HDDs to redundancy. if you place the entire 16-disk set in a raidz2/RAI6 (which is the bare minimum I would call recoverable), your net space is then reduced to 280TB.
So let's say your colocation costs just $100/month, and you have no other costs (ie, nothing ever fails) for a whole year -- that would mean a total cost of 5000 + 100*12 = $6200 / 12 = $516/mo for these 280TB, which boils down to ~ $1.85/TB/mo.
So yes, it's "under" $2/TB/mo, but less than 8% -- and given the initial investment plus all the labor and aggravation this is certainly going to generate, I'd rather pay $2/TB/mo to iDrive e2 S3 as @kapitainsky has pointed out and let them deal with all the hardware/redundancy/energy/failures/etc.
If you found a significant lower number than $1.85/TB/mo, I would be grateful if you could share the details with us.
iDrive e2 S3 is actually $1.67 - but you have to pay per year and commit to some capacity - so effective cost can be a bit more - depending how well you plan it.
I saw the page you linked, but it's not clear it's $3.92 for the entire plan (ie all 8 users) or for each user. I tried signing up for a test ($8/mo would still be cheap) but the effers want my phone number, no email option -- and I don't give my phone number to just about anyone. There's also the matter of how well it works (or fails to), I see rclone has an Yandex Disk remote, but haven't heard from anyone using it. Anyway, if it's indeed $3.92/8TB/mo, we would be under $0.50/TB/mo which is a really good price, even if one has to get multiple accounts and pool them all under a union for >8TB total storage.
You are right -- but given all the rug pulls and shenanigans we've been seeing, even from previously very reputable names like Dropbox, I'd be really wary of paying for an entire year up-front to anyone. And iDrive is kind of an unknown for me.
Case in point, I was very tempted to pay $15/mo as an yearly payment instead of $20/mo for dropbox -- a 25% discount is very considerable and even larger than the one iDrive is giving. But now I'm really happy I opted for the monthly option. There's no guarantee a similar situation won't happen with any provider, so I would recommend anyone to start with the monthly payments and switch later to annual if the provider gives you that option.
You are right. Everybody has to exercise own common sense and do due diligence.
BTW - they are old (older than dropbox) well established company but before mostly known for computers online backups. Their S3 offering is new stuff.
For me difference is fundamental when I see pricing model. with iDrive (as with AWS/B2/Wasabi) I can see that does not matter how much business is coming their way it is sustainable - and probably something they want. All "unlimited" or weirdly cheap offerings ring alarm bells - these guys want business but not too much usage. As too much usage will kill their "business model" - many customers paying and hopefully not using much resources. And even some are probably close to scam territory - and we will see more of these soon as people always like to believe in unbelievable:)
Point taken. But I think every vendor, even with metered storage, is happy in charging a price under their real cost as long as it brings business their way, and betting that many people won't use their full allotment anyway (eg, @Hoptional paying for 400 TB but using only 320TB -- simply because there's no option with exactly as much storage as he needs). Some time in the future they will review that and 'adjust' things -- and I bet the adjustment (throttles, limits, changing rules, etc) will be against the user interests.
After all, if they just wanted/needed the money in advance, a solid company can get a one-year money loan for way less than the 20-25% discounts we are seeing -- so something else must be afoot.
That's my rationale for not giving them any more of my money upfront than I actually have to.
And even if there aren't any right now, I would bet there eventually will -- as soon as the storage vendor gets wise to it.
It really isn't, but there's a personal plan offering 1TB for $2.42, it wouldn't make much sense to charge $3.92 for 1TB per user in a combo. Also if it's charged per user and it's higher than the individual plan there's no reason to set a limit for the amount of users.
I bet they simply expect most families won't register all 8 users or that most of them wouldn't fill much of the available space. That's why there's an individual plan charging $6 for 3TB because they don't expect a single person to be comfortable splitting storage across multiple 1TB accounts.
This is also in line with "family" plans from other services I know, you don't pay per user, you buy the plan and can share with up to n users at no additional cost. Microsoft 365 Family is very similar, but up to 6 users, so 6TB.
There's also this in FAQ:
How do I add new people to Yandex 360 Premium Family?
Go to the Family page in Yandex ID. Share the invitation with your family members by email, phone, or link.
I believe they would say something about purchasing for new users if that were the case.
One small problem is that Yandex is russian company doing business in places like Iran. Some people do not care but for many it can be no go - also for business reasons.
Thanks for the response, and I agree, it makes much more sense if it's $3.92 total (for all 8TB) per month.
So we need a volunteer to sign up and test; if anyone does, please report the results here.
Ha ha, you are right - certainly not a game everyone bothers to play
You are of course right. Morals apart, one could also add that Russia is an unstable country, where laws and rules don't mean much, and to top it off it's currently involved in a war it's not winning...
Exactly. This kind of situation is like an arms race -- the users looking for loopholes in the rules and the provider changing the rules and/or enforcing them better. In the end everyone loses IMHO.
I do not want to brings politics here:) But there are often legal obligations who you can do business with and where you can store data - especially when we are talking about anything sensitive - like personal info etc. One would not like to be on receiving end in case of some data breach to explain why data was handled by company in russia:) it was cheap might not be the best defence
Ha ha, you could be right - and if losing is this much fun, I don't mind losing at all