INEFFICIENT DATA STORAGE SOFTWARE
PROBLEMS
Fixed hardware limits
Data storage controllers or servers have hard resource constraints in CPU cycles, memory, and internal bandwidth. Most data storage systems are limited to dual active-active or active-passive storage controllers.
Each storage controller or server is normally dual socket (2 CPUs). Data storage software is layered over years reusing as much code and open source as doable. Part of this comes from business pressure to get to market as soon as possible. As new features are added, so are more layers of code, and more inefficiencies. No software developer ever wants to redevelop their code, which is what they have to do to make it efficient.
Data storage software code was developed under the influence of Moore’s law when resources and performance doubled every two years.
That’s not happening anymore. Moore’s law has slowed to a crawl. That inefficient storage software results in much fewer resources for IO.
Max IOPS and throughput don’t add up
As previously discussed in the performance section, the data storage software inefficiencies become clearly evident when comparing maximum data storage system performance to the max performance of all the drives at max capacity.
INEFFICIENT DATA STORAGE SOFTWARE
WORKAROUNDS
Common Decidedly Inefficient Software Workarounds:
Limited, Costly, and/or Unsustainable
Buy more data storage systems
As nutty as this sounds, this is the most common workaround for your inefficient data storage software. It adds a lot more complexity and cost.
Complexity in determining which of your workloads go to which of your storage systems.
This requires you to have considerable expertise, knowledge, and time. Ok for a handful of workloads, but hundreds or thousands, not so much.
Management of multiple storage systems for performance is not linear. It adds much more than 100% per system for you.
It increases your management work minimally geometrically and more likely exponentially. Too many redundant labor intensive tasks.
We humans are not very good at redundant management tasks. Some workloads will require you to load balance your multiple storage systems.
Troubleshooting will also take a lot more time and effort.
Buy a multi-data-storage-controller/server system that goes beyond 2
Also known as ‘scale-out’. More controllers or nodes do not cure your data storage inefficiencies. They just add more data storage compute resources at a considerably higher cost. And there’s diminishing marginal returns for each node.
INEFFICIENT DATA STORAGE SOFTWARE
STORONE SOLUTIONS
Unparalleled software efficiencies
Written from scratch for over a decade, the StorONE Engine is specifically architected to use fewer data storage server/controller resources.
No other data storage software comes close to the StorONE Engine’s efficiency. This becomes obvious to you from the zero impact on performance for vSnap snapshots, vReplicate replications, and vSecure encryption.
The StorONE Engine leverages the common metadata table to orchestrate these data services to interact with IO highly efficiently, allowing you nearly limitless use.
Your ability to get increased performance from the same hardware and media is additional evidence of that unique efficiency.
Virtual Storage Container™
The incredible efficiencies of the StorONE Storage Engine enable you to create separate block, file, and object volumes completely multitenant.
Each Virtual Storage Container™ has its own self-contained customizable storage services while providing you the ability to share media between them seamlessly.
