Storage Informer
Storage Informer

Storage Made Simple?

by on May.26, 2009, under Storage

Storage Made Simple?

EMC logo Storage Made Simple?

Ever have to explain this storage stuff to someone who doesn’t really have a background in it? I find myself speaking to all sorts of people these days, including more and more that are relatively new to all of this….

Ever have to explain this storage stuff to someone who doesn’t really have a background in it?

I find myself speaking to all sorts of people these days, including more and more that are relatively new to all of this.   One example is at press events — there are people who follow these topics closely, and those that are relative newcomers.

The first few times I tried to explain the wide world of storage to an outsider, I’m afraid I made a terrible mess of everything, and ended up leaving people more confused as a result!

As a result, I’ve been searching for a quick way to explain to people why storage is such an interesting topic to so many people.  I’ve sketched out some of the key ideas below.

Let me know how I did?

It Starts With Information

You can’t really understand why storage is interesting unless you have an appreciation for information itself.

Our society is generating great enormous gobs of the stuff at an accelerating rate — so much so that we have to make up new words like “zettabyte” to just describe how much information we’re all producing.

Not all of that information is useful or valuable, but if it is — it needs to be stored somehow, otherwise it’s largely lost forever.  And information can be just about everything — from your home movies to your health records.

Storing information implies that we might want it back again at some point in the future. 

And that’s where the fun starts.

Storage Is A Commodity, Information Is Not

Many people who aren’t close to this topic assume that storage devices are pretty much like other IT technologies: processors, memory, networks, etc. 

While it’s true that all technology components are essentially commodities, I think storage is subtly yet profoundly different — unlike other technologies, it stores information, which has value.

For example, the CPU on my laptop can fail.  No problem, I replace the CPU, and all I suffer is a bit of inconvenience.

By comparison, should my disk drive fail (and I haven’t backed up the information), I will not only have to replace the disk drive, I will likely lose hundreds of gigabytes of my personal information. 

That’s more than just an inconvenience to most people.

The Storage Stack

To encompass a broader storage discussion, we have to go beyond just talking about disk drives, for example.

One model that I tend to use has four layers:

storage media: disk, flash, tape, FC, SATA, SAS etc.
storage architecture(s): mid-tier, enterprise, clustered, NAS, SAN, DAS etc.
storage functionality: tiering, replication, deduplication, RAID, management, etc.
business outcome: cost, risk, value

Sure, there’s always more to talk about, but it turns out that you can hang just about any storage discussion around these four elements.

Storage Media

Right now, the storage industry is getting exposed to an entirely new storage media.  For many years we’ve lived with tape (somewhat on the decline) and different flavors of disk drives (used for most storage applications).

But as of late, we’re seeing more and more solid state storage enter the marketplace.  Solid state memory doesn’t use rotating platters, it uses memory chips to store information.

You’re probably familiar with the consumer applications of things like flash memory in USB sticks and music players.  A different kind of flash memory is used in enterprise IT environments, but it’s catching on there as well.

All these different kinds of storage media represent tradeoffs in cost, capacity and performance.  For example, tape is very cheap and relatively big, but painfully slow for many applications.  Enterprise flash is very fast and energy efficient, but relatively expensive from a capacity perspective.  And disk drives are somewhere in between those two extremes.

Over time, storage media keeps getting cheaper and cheaper — but it seems that we’re generating information at a faster rate than prices are declining!

The trick is to get the right information on the right storage media at the right time. 

Storage media can’t do that by itself — it needs external software (storage funtionality) to decide which bits of information need to be stored on specific devices at a specific time.

Storage Architecture(s)

At some point, you’ll need more than one storage device such as a disk drive, which brings up the question of how multiple storage devices can be assembled into something larger — often called a storage array or storage platform. 

We’ve now left the world of storage devices, and entered the vast universe of storage architectures.

Engineers take storage devices, and combine them with other computing ingredients (processors, memory and CPUs) to build a wild world of storage platforms ranging from the modest to the magnificent.

All of these different combinations represent tradeoffs and compromises — there is no one “best” architecture for storage, despite what vendors might want you to believe.

Most enterprise IT organizations have a wide variety of storage requirements.  Rather than having dozens of different storage platforms for each need, the tendency is to standardize on a small handful of storage platform types that cover the range of requirements.

New requirements evolve in enterprise IT storage environments all the time. 

For example, there was no need for storage that was good at, say, storing archived emails until we all started getting deluged with emails. So one source of innovation around storage architectures is coming up with new approaches to relatively new storage problems.

The other source of architectural innovation comes from the components themselves, as the underlying technologies are evolving rapidly (storage media, processors, memory and network connectivity). 

This means that — generally speaking — newer storage architectures are being continually created that do a better job of solving customer problems than traditional approaches.

Storage Functionality

Build a storage array or storage platform that has storage media, processor, memory and network connectivity, and you’ve essentially build a special-purpose computer.  And all computers need an operating system and applications to do anything useful.

Storage arrays all have operating systems.  They’re often referred to as “microcode” or “array software”, but the idea is the same as an operating system.  And these storage operating systems support storage-specific applications (also called functionality) that do useful things.

One of the first useful things that storage arrays learned how to do is to protect against a disk drive failure.  They’d make redundant copies of information on multiple disk drives (also called RAID) so that if you lost a disk drive, you could rebuild its information from the surviving copies.

Another useful thing that storage arrays learned to do early on is to make a copy of information locally or remotely, often called “replication” or “snaps”.  Just like making a copy of information on paper has all sorts of practical uses, so does making an additional copy of information sitting on a disk drive.

Today, there’s all sorts of useful storage applications that can run on a storage array.  There’s software that can auto-tier information between different media types which can boost performance and save money.  There’s software that can find and eliminate redundant information — referred to as data deduplication or “dedupe” for short — that can save on storage costs.

And that’s just the beginning — management software, archiving software, backup software, search software, anti-virus software — the list of potential storage functionality gets quite long.

It’s important to point out that these same capabilities could be run on a normal server, or perhaps a device sitting in the storage network.  Indeed, for a given piece of software functionality, it’s often the case that it can be done in any of three places, and not entirely unique to storage arrays.

This inevitably brings up the question of “where’s the best place to run storage functionality” which usually results in the inevitable answer: “it depends“.

And, as of late, there’s renewed interest in having someone else provide storage services and functionality on your behalf without having to actually own the underlying technology.  This idea of “storage as a service” is often one part of a broader “cloud” discussion.

Business Outcome

As a final stop in our journey, it’s important that we step away from the technology, and look at what individuals are trying to get done.

Back at the beginning of this piece, I said it’s all being driven by the need to store information for later use.  In a business context, someone decides that something is worth keeping, so it’s stored until later.

But, of course, storing information isn’t free — which brings up a whole discussion around minimizing costs and charging people for what they use. 

Most organizations want to minimize costs.

And storing information entails a certain risk to the business — the risk that it’s not there when you need it, or that sensitive information is secured, or that compliance rules are followed. 

Most organizations want to minimize risks at the same time.

However, we can’t forget that information also delivers business value.  For example, if it takes customers 5 minutes to complete a web form, you’ll hear about that — if you have any customers left, that is!  And a business analyst trying to figure out what’s going on in the business might want their information in minutes rather than days. 

Most organizations have to consider the value side of the equation.

Cost, Risk, Value — the question becomes how do we dynamically balance the outcome as business requirements – and technology — constantly change?

Let’s not forget — it looks like that every year, we’re going to be generating far more information than the year before — forcing a continual re-examination of the question “how do we best store it all?”

And that is what makes storage so interesting to many of us. 

Update your feed preferences

URL: http://emcfeeds.emc.com/rsrc/link/_/storage_made_simple__90573164?f=84f8d580-01de-11de-22d1-00001a1a9134

:, , , , , , , , , , , , , , , , , , , , , , , , , , ,

Leave a Reply

Powered by WP Hashcash

Looking for something?

Use the form below to search the site:

Still not finding what you're looking for? Drop a comment on a post or contact us so we can take care of it!

Visit our friends!

A few highly recommended friends...