Hold it! Don't back up to a cloud until you've eyed up these figures
Trevor Pott drills into the real price of online storage
A terabyte is sod all. What about moving 15TB a month?
I also have one customer cheerfully ploughing through approximately 15TB per month of backups. Now, 31 Amazon EC2 volumes at 500 gigabytes each transferring in 15TB of data costs $1547/month. Assuming I take the $900/month fee the ISP charges just to light the fibre up out of the equation, we pay $2,800/month for a sustained 50Mbps connection, no metering. Assuming 30 days in a month, this has a theoretical capacity of approximately 15.45TB a month. This lines up with what I see in the real world; we flatten it all month long and just get the backups done. This makes the monthly cost of the solution $4,374.
The 4TB 7200 RPM Hitachi Deskstar sells for $329 at my local computer retailer. Five of these drives (for RAID 5) is $1,645; a Synology DS1512+ costs $899. A 10x10 storage unit is $233/month, and the delivery guy costs me $33 per run. So for me to back up 15TB off-site each month is $2,800 per month.
Any time we wanted to do a recovery of that data, Amazon would charge us $1,689.51. I would also have to up the bandwidth allocation for that month by another 50Mbps, meaning an additional $2,800. This makes the cost of recovery $4,489.51, and it would take a month to pull down the data. Even if I wanted to, I can't make it go faster; the fibre pipe the ISP can provision me only goes to 100Mbps.
Everything about those calculations is wrong
The above needs to be taken with a bag full of salt. These are back-of-beer-mat calculations that leave out a lot of important considerations. At first blush, 1TB of online storage doesn't compare too badly with the local delivery guy's solution. The 15TB one is obviously nowhere near feasible.
This misses a couple of things. The online solution presumes that my 1TB of online storage is completely overwritten each month. If not, (as is the case with many types of backups) then the cost of storage will rise perpetually. The local storage option also presumes that the backup "tapes" are purchased and then never re-used. In some backup schemes, this is true. In others, these "tapes" are rotated out, being overwritten again on a regular cycle. I tend to use one where there are a total of 12 "tapes" in play; backups occur weekly, and each "tape" is recycled every three months.
The cost of connectivity is also highly variable. If you are sitting on an internet pipe so fat that bandwidth is measured in cents per gigabyte, then cloud storage economics shift entirely. What seems economically reckless to the small and medium enterprise market starts to make a lot more sense at scale.
What about legal jurisdictions?
I have a massive DVD and Blu-ray collection. I have ripped every single one of these discs to my NAS. Until just recently, this was legal in Canada; no "digital locks" laws had yet been passed. Those ripped movies are perfectly legal for me to retain; what is their legal status if they are synchronised to a cloud service? If they are synchronised to a service in a country where their mere creation is illegal (because it required breaking encryption) can I be financially ruined for backing them up?
What about an ebook of George Orwell's 1984? In Canada, Australia and other nations this novel is a public-domain work. It is not public domain in the United States. Assuming the ebook I am using has disclaimed all rights (there are several such works), or I have scanned a print version myself, under what laws am I judged if this gets stored on an American server?
What if it merely travels through an American router or American-owned fibre pipe? Corporate liability regarding intellectual property remains legally nebulous in many places.
We can do this cloud storage thing, if we choose to. In some circumstances, it makes good business sense to do so. In others, it would be ridiculous to even attempt.
Setting aside legal ambiguities, the limiting costs of cloud storage aren't the cost of the service provided by companies such as Amazon. Cloud storage providers offer storage cheaper than most companies can source it themselves. The boundaries on cloud storage are the throughput and rates provided to us by our ISPs. The cloud storage revolution can't truly begin until the cost of internet access comes down. ®