Key Criteria for Evaluating Primary Storagev2.0

1. Summary

Until recently, mission- and business-critical functions in large enterprises were concentrated in a few monolithic applications based on traditional relational databases. In this scenario, block storage was often synonymous with primary storage—that is, storage dedicated to holding an organization’s most critical data, which was usually structured data, such as a database. Primary storage therefore required the highest levels of performance, reliability, and availability, and most solutions were designed for exactly that. As a result, performance, availability, and resiliency were generally excellent, but these qualities often came at the expense of other characteristics, like flexibility, ease of use, and cost.

Ideas around primary storage, data, and workloads have changed radically in recent years, due to new technologies, architectures, and development models on the one hand, and to the realization that data is one of the most important assets for any type of organization, on the other. Data is no longer coarsely categorized as structured or unstructured, but instead by the applications and business processes accessing it. In fact, thanks to digital transformation and other initiatives, it is now common for applications to access structured and unstructured data concurrently, with different protocols, based on separate logical or physical repositories.

When it comes to modern storage in general, and block storage in particular, flash memory and high-speed Ethernet networks have commoditized performance and reduced costs, allowing for greater flexibility in system design. At the same time, enterprise organizations want to align storage with the rest of their infrastructure strategy, based on pressing requirements, including:

• A more agile infrastructure that can respond quickly to business needs
• Improved data mobility and integration with the cloud
• Support for larger numbers of applications and workloads concurrently on the same system
• Overall infrastructure simplification
• Automation and orchestration that speed up and scale operations
• Better infrastructure performance that is reflected in user or customer experience
• Drastic reduction in the total cost of ownership (TCO) along with a significant increase in the capacity under management per sysadmin.
• Cloud-like consumption models

Many of these points are crucial to the realization of modern IT infrastructures and strategies and have contributed to the increasing number of storage solutions available from both startups and established vendors. Traditional high-end and mid-range storage arrays have been joined by software-defined and specialized solutions serving similar market segments, but differentiated by their focus on specific requirements. A one-size-fits-all solution doesn’t exist.

Storage solutions have been affected also by the steady increase of public cloud adoption. Paradoxically, while it might appear that the data center would grow due to huge demand for compute and storage resources, organizations seem to trust public cloud providers’ abilities to deliver enterprise-grade services, resulting in more workloads being shifted to the cloud. This broader adoption of cloud services makes customers more inclined to adopt cloud-like, OpEx-based consumption models. Thus vendors have started proposing Storage-as-a-Service (StaaS) as a critical way to differentiate from their competitors, offering customers more flexible consumption options.

In this report, we analyze the important features of modern storage systems to see how well they respond to enterprise needs and to enable organizations to evaluate specific solutions based on their own requirements.