What’s in a Nine?
The recent revelation that Ceph-based object, file and block stores at SourceForge and DreamObjects have been knocked out has led me to ponder the HGST Active Archive 15 9’s data durability statement, feature and promise.
It appears that both SourceForge and DreamObjects suffered from system downtime. This is a total lack of system availability. Both also suffered from data durability. That is the management of bringing current data back online along with the systems. Without a clean data store, the management and reconstruction of data-at-rest, as well as any data rebuilds, add to the frustration of bringing systems back live. System reliability and/or availability represent just one factor when determining the risk and gains when deploying storage. Durability of the data should be considered as strongly as availability and system reliability.
The difference between durability and availability is fairly simple. Durability is about what happens when all power goes out everywhere. Has all data been written to stable storage that doesn’t require power (e.g. disk/flash), in a form that allows it to be brought back and used? Availability is about what happens when there’s a partial failure – a disk, a node, a network. Does the system continue to operate and provide the same services it originally did?
HGST 15 9’s Data Durability in Perspective
99.9999999999999 = 15 nines
That’s a whole lot of nines. Most people don’t realize that durability is an issue in environments. Traditional IT storage admins have had to contend with terabytes or even petabytes, but when you’re a cloud provider, you’re thinking exabytes and zettabytes. When put into Cloud IT vernacular, it means that the HGST Active Archive System can deliver, thru erasure coding, (over the course of multiple years) clean, coherent data in the event of a system-wide power loss, or unplanned downtime. That’s a whole lotta data durability.
Ok. Pretty cool stuff. But what does that mean for my business?
High availability has been defined by the “nines” metric for years and years. System-level architectures have been designed around reducing downtime, failing in place technologies as well as heroic recovery algorithms to meet the demanding SLA’s of business line owners, IT leaders and ultimately CIOs. As businesses realize and monetize the valuable data at hand, this durability promise becomes more important and more easily understood.
The Cost of Downtime
A 2013 study by Ponemon Institute claims the cost to be approximately US $7,900/minute.
“The study of US data centers quantifies the average cost of an unplanned data center outage at slightly more than US$7,900 per minute. This is a 41% increase from the $5,600 it was in 2010, when Emerson and Ponemon first started downtime cost assessment studies.”
Applying the results of the study to the two outages claimed above with Ceph would have huge monetary implications, depending on total outage from SourceForge. Perhaps even as, if not more, significantly, there is the loss of trust and vendor selection associated with severe business outages.
HGST Active Archive System Provides Unprecedented Value
Reduce risk, increase productivity, and deploy the HGST Active Archive System storage solution. Deliver enterprise cloud-scale economics with the performance, security and flexibility required for on-premise data centers including unbreakable durability at greater than 15 nines and non-disruptive scalability. Do all this and forget the headaches with data availability faster than a blink of an eye.