Our service is unique in that we build a highly scalable cloud
object storage, using existing open-source software as much as possible,
in order to cut the development and support costs. This allows us to
build a system core very quickly with minimal investment.
Our system design uses commodity server hardware. It is
fault-tolerant and self-healing. This makes the S3for.Me system very
cheap compared to enterprise storage arrays. It is built for the Web and
it can serve all objects directly using HTTP or HTTPS protocols, which
makes it the ideal choice for modern applications.
Every object is stored on three different physical servers to
ensure very high object reliability. The annualized failure rate (AFR)
for the HDDs that we use is less than 1%. This means that the AFR for 3x
redundancy installation has an AFR equal to 0.01 * 0.01 * 0.01 =
0.000001. Consequently, we can guarantee a 100% - 0.0001% = 99.9999%
object durability, which is considerably higher than most other storage
Our data center provider guarantees an annual average of 99%
network availability for the data center network infrastructure. We use
multiple data center deployment and our servers are installed at least
in two data centers at the same time with full data copies at each data
center. This means that we can guarantee 100 - (0.01 * 0.01)*100 =
99.99% object availability over a given year.
The access control list (ACL) system we use allows for two types of objects:
- Public - anyone can access it with the link.
- Private - a valid key is required to access it.
And there are two types of keys:
- Permanent - Used to access and manipulate bucket objects; offers full control and no time limitations.
- Temporary - Used to access objects only and has an expiration date.
We use the most popular Cloud Storage API to date - Amazon S3
which makes migration smooth and easy. You can use any software or
library to work with our service. If you have any problems or questions,
our support team will be glad to assist you any time.