In part one of this series we looked at how storage technologies play an important role in 21st century human activity. As we progress further through the series we will look at more complicated subjects, along with some elegant, and not so elegant, architectural approaches to recognize and possibly avoid where possible. As for now, lets cover off some of the building blocks of storage architecture.
By Glenn Augustus, Technologist, Global Infrastructure Services
To start with let’s look at disk (or disc, they are interchangeable) – it’s what people generally think about when storage is mentioned, whether it’s a disk in your PC at home, your iPad on the move, or the enterprise storage array for the business critical application, the principle is the same, to store a copy of data that is persistent and easily accessible in a random fashion. The form it takes depends on the type of storage but it’s usually one of two types, one or more mechanical spinning disks (known as platters) combined to create a large capacity hard drive, or more recently the non-mechanical solid state storage built on a silicon substrate, commonly referred to as Flash or SSD, where performance and support for intensity of workload outweigh some limitations in potential lifespan. These issues are a little deep for the basics but suffice to say there is more to the deployment decisions than cost per GB, one of the subjects for a later post.
Tape is an aging technology, consisting of a magnetic medium on a plastic film, passing a read/write head to support primarily sequential access to data, as you can imagine random access is more of a problem for tape as the time to access the data (seek time) is long in real terms (tens of seconds), there is a term that you will hear some of the storage folk use called ‘shoe-shining’ this is where the tape is moving backwards and forwards more than usual to service the requests. Tape is being rapidly encroached on by disk solutions with built-in intelligence. However, the simplicity of tape keeps it at the table and plenty of deployments inside CSC, which testifies to its continued use. There may still be cases where a new tape solution is appropriate, but these are becoming increasingly rare.
Storage Arrays are the apartment blocks of data housing, providing a high concentration of capacity and performance in a comparatively low footprint, they are pre-plumbed to be able to connect to and provide a range of storage services. Most modern arrays go way beyond their traditional functional sweet-spot of delivering blocks of storage through open standard or proprietary connectivity solutions – these new arrays are known as unified arrays, as they ‘unify’ what would require multiple devices into one to offer such things as network attached storage, wide area replication and block storage services.
The basics continue with some easy explanations of terms like Volume Manager, File System, Capacity and Latency
Don’t forget to comment, get involved and get in touch using the boxes below and reach out direct @glennaugustus
As a Technologist in CSC’s Global Infrastructure Services, Glenn helps clients use technology to realise effective IT through the development of CSC’s infrastructure services portfolio. He has held a variety of senior architecture and engineering positions within CSC before becoming Global Offering Manager for CSC’s Storage as a Service and most recently Chief Technologist for Compute. Glenn lives with his family in the United Kingdom.