Enterprise class storage without the cost and complexity
Storage systems have become extraordinarily complex, expensive and highly proprietary
The basic architecture for current storage systems was designed about two decades ago.
These systems were designed primarily for high volume transactional processing and they are good at it.
The last major storage innovation was Network Attached Storage introduced over a decade ago.
This simplified some things but came with limitations.
Now enterprise class NAS has followed the same direction and evolved into large,
complex platforms with proprietary software stacks from top to bottom.
Storage is consumed like a commodity, but storage systems are anything but.
With the rate of data growth at close to 60-100% CAGR , storage systems have
had to get much bigger than originally designed for while the underlying architecture has largely been unchanged.
The relentless growth in data has forced storage systems designers to packed more and more processing, networking, disks,
controllers and many proprietary technologies into these boxes to enable them to meet this new reality.
The Result
Storage systems have become very complex requiring highly trained storage engineers to configure and manage.
High system costs due to the high end components required and complex manufacturing even while the actual cost of storage drops 30% year on year.
Business models that are only profitable on 60% gross margins.
An industry that looks more like the mainframe industry of the 80's with customer lock-in and providing everything from men in white coats to operating systems and critical applications that could be purchased only from that vendor.
We asked a simple question - why cant you buy storage like a commodity?
After all, the disk manufactures have made extraordinary advances in the capacities
and speed of hard disks while driving down the cost 30% per year for the past decade or more.
The answer was that storage is complex - it is anything but a commodity.
If you want a system with scalable capacity that is highly reliable and with high performance - it gets more and more complex
(and costly) to design these systems. Like most things that evolve over time and try to meet new requirements as they go along,
you probably wouldn't start there if you wanted to get here.
So we took a step back and asked ourselves - Is there a better place to start from?
The answer was simple once you asked the question. Eliminate the complexity -
get rid of the big complex systems approach. There is no need for it.
Replace the big complex systems with small, simple building blocks of storage.
And replace the big, complex storage processing platforms by harnessing the incredible and ever increasing
processing power that sits idle 90% of the time in powerful PC's to become the processing engine.
The combination of the two is a building block model for storage using simple
and affordable storage blocks, that can be added one at a time to a network by anyone when they need more capacity.
It is also an incredibly scalable model as the numbers of machines that access the storage grows, every machine
adds more processing power to the gridstore. 25 modern PCs contain more processing power than some of the
largest enterprise storage platforms.
Your storage processing power effectively grows at zero cost (you already paid for it).
So where does this road lead to?
The current demand for storage capacity is just scratching the surface now.
From here, practically everything in our lives will be digitized in some shape or form - and stored
somewhere practically forever.
Our vision for storage is as much about the industry as it is about technology.
To meet the demand for vast quantities of storage that is simple, affordable and reliable - both need to change.
Gridstore represents a new model for storage and a new model for the storage industry.
Open Industry Model - Change the industry from what looks like the 80's mainframe industry,
to something more resembling the PC industry today built around high volume manufacturing and global distribution.
Profitable OEM Partners - Producing storage nodes in high volumes to meet the exploding demand.
While the simplicity of the model will remain, storage nodes will come in different forms to meet different demands.
OEMs with high volumes can be profitable with 15% gross margins instead of 60% like today's large storage vendors.
Choice for Customers - With multiple storage node vendors to chose what best fits their needs.
From small grids for file serving and backup to large scale transaction processing and cloud computing environments -
NASg is a scalable platform that delivers low cost, highly reliable storage for different scenarios.
Eliminate Waste - We purposely set out to build a platform that could leverage existing storage resources.
The waste of the "rip and replace" upgrade is absurd in todays economy and the future. NASg allows you to continue using resources
until they need to be replaced - and then makes it simple to do this.
Minimal Carbon Footprint - By leverages the existing compute infrastructure for processing power,
NASg eliminates a significant proportion of processing platform from the storage infrastructure.
It also allows low powered devices to work efficiently in the middle using a fraction of the energy and cooling common with large systems.
This translates into low cost storage, low power consumption and a reduced carbon footprint.
|
