Self-Storage (a la BackBlaze pods) vs Amazon S3

wellman

I've been scratching my head as I read BackBlaze's latest post on "Petabytes on a Budget v2.0." With a dataset as large as SmugMug's, how can outsourcing the storage to S3 be economical? You all certainly know your business and your craft, so I must be missing something.

BradfordBenn

I cannot speak for SmugMug, but one of the things to consider is the data durability and redundancy. That price is shown as one Pod. So if you want one in another location and to mirror it for redundancy you have to double the price, plus include a high speed data connection between the two. Then if you are charging people for the storage, speed is going to be a key issue. I am not willing to have slow speed for that connection if I have a 30Mbps pipe to the Internet I expect my backup location to receive at 30Mbps. Now lets say I have multiple clients, all will be expecting that same speed. So you need a very big pipe from your data center to the Internet. Plus you need load balancers and other items to actually manage the data.

So while the hardware cost is fairly low to do a pod, that is the cost of a single pod. It is not the cost of getting data to and from it nor getting the redundancy. If my business model required good speed and reliability I would be willing to pay for that level of support. I also have a hunch that if you are buying a Petabyte of data storage there might be some volume discounts.

denisegoldberg

wellman wrote: »

With a dataset as large as SmugMug's, how can outsourcing the storage to S3 be economical? You all certainly know your business and your craft, so I must be missing something.

I don't think you're comparing like things here. As far as I know, photos are served from S3 - it's not just a backup. The other services you noted are backup services.

Do a search for smugmug amazon s3 to see some of the articles / information.

--- Denise

Andy

http://twitter.com/#!/DonMacAskill/status/93796872778563584

W.W. Webster

wellman wrote: »

With a dataset as large as SmugMug's, how can outsourcing the storage to S3 be economical?

You reckon SmugMug pays retail?

Dan7312

Amazon S3 storage is replicated to three different, geographically spead out sites which, all else being equal, increases the reliablity of the storage. Amazon also offers a less expensive option with only replicates the data over two sites.

For Backblaze, at least according to their site, all the data is stored in a single location. Also backblaze is a mirror of what is on your computer. If you delete a file on you computer it will eventually be deleted on Backblaze.

Each has it's own uses, but they are a bit different from each other.

wellman wrote: »

I've been scratching my head as I read BackBlaze's latest post on "Petabytes on a Budget v2.0." With a dataset as large as SmugMug's, how can outsourcing the storage to S3 be economical? You all certainly know your business and your craft, so I must be missing something.