weka blog detail third level banner

What Does It Really Mean to Be a Software Storage Company?

October 1, 2019
Liran Zvibel
What Does It Really Mean to Be a Software Storage Company?

Liran Zvibel, Co-Founder and CEO of WekaIO, reflects on the question “What does it really mean to be a software storage company?”

Storage products moved from being hardware-centric to software-centric quite some time ago.  My previous company, XIV Storage (which was acquired by IBM), did not even have a single hardware designer.  XIV Storage was centered around highly differentiated software layered on top of Commercial-Off-The-Shelf (COTS) hardware, as was Isilon, which came to market in the same era. Yet neither company was truly a software storage company.

What does it actually mean to be a software storage company?  There are several essential elements:

  • no reliance on specific hardware configurations or models
  • the ability to work on standard servers from any OEM vendor (such as HPE, Dell, Cisco, Lenovo, Supermicro, Inspur, etc.)
  • no reliance on storage vendors for hardware logistics
  • capability to run the same software on-premises and in the public cloud
  • an OPEX-focused pricing model based on software subscriptions, unlike the perpetual (CAPEX) licensing of appliance-based systems.

Back in the hard disk drive (HDD) days, enterprise storage systems went through a rigorous burn-in process during manufacturing to weed out bad drives and ensure that customers were getting systems that were reliable and performed to specifications. Flash-based storage has dramatically changed the production process as there are no moving parts and there is little variance between the different units. The elimination of a lengthy burn-in process to improve product reliability is a key factor in allowing the implementation of a software-based storage model.

The traditional approach for storage vendors has been to sell products with a “perpetual” license that lasts 3 to 4 years, after which support charges became so high that customers are forced to do a forklift upgrade. If you look up the word “traditional” in the dictionary, its definition states that it means “never-ending or never-changing”. Why has the industry settled on this?  There are two reasons for this traditional approach. The first is that originally, storage systems were HDD based. Every HDD has a forklift replacement in its future, as its warranty lasts only for 3 to 6 years, hence traditional storage vendors do not provide longer warranties for their systems. Secondly, first- and second-generation flash-based systems did not scale or run mixed workloads very well. For those reasons, customers end up buying a lot of small, siloed storage systems. Each of these silos runs on its own hardware, and when that hardware ages, it will have to be replaced.

A software-only, hardware-agnostic storage company is not only good for customers, but also for storage vendors

  • customers are empowered to choose their hardware vendor rather than be limited by storage vendor lock-in (envision running IBM storage on Cisco hardware or NetApp storage on Dell EMC hardware); the customer can leverage larger server purchase contracts to get the best overall storage pricing.
  • customers can choose the software storage solution they want and plan a deployment model based on their budget. For example, they may wish to deploy the software on commodity white-box for their sand-box or staging environment while purchasing fully pre-configured brand-named solutions for their production environment.  These options are now a reality with a software storage model.
  • customers can run the same workloads on-premises and in the cloud, simplifying operations and making hybrid cloud more real
  • the customer no longer suffers the ongoing pressure of the 3-4 year forced upgrade cycle
  • the storage vendor can now employ an agile software release cycle with continuous feature integration on a monthly basis instead of the yearly forklift upgrades
  • the storage vendor benefits from a software-only model because the business is not burdened with the cost of inventory, stocking logistics, shipping, and hardware maintenance.

Software-based storage products that run on commodity hardware and are scalable enable customers to easily change the size of the deployment over time. Customers don’t have to guess the size of a four-year solution on day one and be stuck with it for “perpetuity” (e.g. 4 years). Imagine a software solution that allows you to consume hardware based on demand rather than on a four-year projection, which will more than likely be wrong. Today’s flash devices and modern servers can hold up for more than 3 or 4 years, allowing customers to keep leveraging their hardware investments for much longer, augmenting them with new hardware purchases over a continuum.

Furthermore, in many cases, hardware-centric storage vendors offer new features only on new hardware platforms as they rely on hardware advances to drive feature differentiation. Take the example of encryption where many vendors depend on self-encrypting drives to deliver encryption at rest. In contrast, software-based storage customers can upgrade their feature sets by simply downloading the latest version, thanks to a software subscription model.

Modern IT organizations are going to increasingly expect and demand software-only storage products because of their many customer advantages. With the leaner and more flexible processes of building software-only products, storage vendors can focus on the core business of delivering the best possible storage feature set and the best value to their customers.

For more information on WekaIO’s approach to storage, read the following architecture white paper here.

close-btn