Heard this before, what does it signify?
First, some definitions are in order, here goes:
Reliability and durability are used interchangeably in storage, and ensures that stored data does not suffer from bit rot, degradation or other corruption. In short, the data you’ve stored, whether 5 seconds ago, or 15 years ago, is still exactly the same data, with no modifications, and nothing missing.
Thats a pretty tall commitment.
Highly reliable storage is important for long-term storage needs, which means the system needs to proactively watch and correct any issues on a regular basis (we call that “Active Integrity Checking”, also known as periodic validation).
Availability (often called uptime), ensures that the storage system is operational data delivered when you request it. A very important point to note is that ; If you have highly reliable storage, but can’t get to it (for example, tape in an offline vault, or a regional/system-wide network or power outage for object store), your stored data effectively doesn’t exist.
Both reliability and availability require robust architectural designs to meet the expectations we all have for storage — that when we want to get to our data we can, and it is exactly the data we expect.
So, whats with the 11 9’s or 5 9’s reliability/ durability?
Fun with numbers
Definition: 99% reliability means that you will lose one object out of 100 every year
So, 5 9’s means and 11 9’s mean in terms of durability ?
Losing 1 object out of 100000 every year and 1 object out of 100000000000 every year. Some math below:
1 – 1/100000
99.999% 1 – 1/100000000000
If durability is the space/object component, then Availability is the time component. Availability if a guarantee of the service, in terms of its uptime.
Availability is how much time the storage guarantee that your data and services are available to you, as in customer. This is usually documented as a percent of time per year, e.g. 99.9% (or three nines) means that your data will be available and will be unable to access the data for no more than about ten minutes per week, or 8.77 hours per year.
So, a downtime of ~17.5 hours per year, is 99.8% availability:
365 * 24 (hours per year)
99.8 / 100 * 8760 (uptime only 99.8 % of time)
== 17.52 (only so much hours per year downtime allowed)
== 17.52 * 60 / 52 = 20.2153846 minutes per week
== 2.887912 per day
Now, thats a huge requirement to meet, however most cloud vendors today are guaranteeing this, both from a durability aspect (space) and an availability aspect (time). However, it’s important to read the SLA for every service that the cloud vendor provides. And after this, factor in the cost.
Thanks for reading.