itinfraworld

Modern Storage Infrastructures for Flash and Cloud Require Intelligent Abstraction

By Gary Quinn, President & CEO, FalconStor Software

Gary Quinn, President & CEO, FalconStor Software

Incorporating any new platform can be complex; incorporating flash and/ or cloud although they are very flexible and advanced in their functions can be a significant challenge due to lack of interoperability and common standards.

"Modernizing the IT infrastructure according to an Intelligent Abstraction model creates a foundation for optimized, simplified, more cost-effective operations"

Like all other service-based, mobile-friendly, software-defined, or virtualized models, modern technologies like flash and cloud deliver greater simplicity, efficiency, and cost of ownership on modern infrastructures. So in order to protect existing investments and merge in new technologies without perpetuating or adding complexity, modernizing the infrastructure itself should be considered first.

Modern storage infrastructures are built on the principles of Intelligent Abstraction. “Intelligent” refers to information movement, management, and protection that is adaptably controlled and “abstraction” refers to de-coupling applications and information from physical hardware. Think of it as a combination of artificial intelligence, virtualization, and policy-based automation, enabled by advancements in storage, server, and networking technologies, to create more cost-effective, agile and efficient systems.

Intelligent Abstraction is not necessarily a product or architecture. It is a fundamental shift in how we view the data center itself: not as a mass of hardware, networks, and protocols, but as aggregated resources or assets that can be provisioned and orchestrated as needed to deliver consistent IT services to the organization. It is an enabling technology that helps simplify and optimize IT.

Whenever storage is refreshed there are some common, but critical, goals:
• Safe, efficient, and cost-effective data migration from old storage to new
• Deployment of required data functions such as snapshots, replication, recovery, high availability, de-duplication, compression
• Tools to manage, monitor, and report on the health, performance, and utilization of storage resources and data services
• Reasonable up-front acquisition cost, acceptable long-term cost of ownership, and predictability of future costs.

With an Intelligent Abstraction data center model, these goals can be achieved across the entire environment, regardless of the underlying hardware, networks, or protocols; regardless of storage type, vendor, or the physical location of the data. Without Intelligent Abstraction, there will be multiple point solutions specific to each hardware platform, vendor, and application.

These islands of storage that are managed separately give rise to duplication of infrastructure and data, an inability to share resources, and different fixes for similar problems.

Software-defined storage is an example or subset of Intelligent Abstraction. Provisioning storage and server assets is simplified, and mundane operational functions are automated by software. The cost and complexity of operating the infrastructure can be significantly reduced by eliminating the mundane individual management of many physical components, and delivering data availability and business continuity across heterogeneous storage and application platforms.

It’s tempting to equate virtualization with Intelligent Abstraction. While virtualization also de-couples applications, data, and workloads from physical hardware to present a consolidated infrastructure, this alone does not perform essential operations like provisioning, data protection, replication, or optimization without underlying software-based intelligence. Virtualization is necessary, but not sufficient, to deliver the full suite of benefits of Intelligent Abstraction which applies to both virtual and physical storage. (However, Intelligent Abstraction should not be confused with so-called hyper-converged architectures that include compute as part of the abstraction.)

Implemented at the network fabric layer, Intelligent abstraction overrides firmware-based intelligence in the storage so data services can be standardized across all tiers and silos. The capability of the storage hardware itself is no longer a restriction, nor is it a consideration when purchasing new storage. Storage instead is a commodity purchase, available from any vendor and licensing costs are eliminated as well.

Flash arrays and cloud storage present particular integration challenges, unless an organization can afford to throw everything out and start over. Moving data to a new and different platform,whether on-premise, remotely located or in the cloud, can be complex, costly, risky, and time consuming. WAN and network bandwidth requirements can soar. Data security procedures, including encryption and key management, must be followed for data in flight. But in the Intelligent Abstraction data center model, all storage is essentially created equal, so applications and data can be moved between locations and unlike infrastructures easily and non-disruptively, at any time, even while applications are live.

No discussion of storage infrastructure would be complete without addressing backup and data protection. Backup is the most fragile and costly component of traditional storage services due to the many moving partsliterally and figuratively. This becomes a non-issue in Intelligent Abstraction thanks to the elimination of batch processes with policy-driven automation, and the ability to provide complete business continuity/disaster recovery horizontally across the data center.

One of the strategic goals of Intelligent Abstraction is to enable cost savings. We have found that the savings can be immediate and dramatic. IDC interviewed eight users that deployed FalconStor software as part of their Intelligent Abstraction initiatives. All customers reported overall IT cost reductions, greater system availability, and increased staff productivity. Users reduced tape storage costs by an average of $162,000 annually and realized an average savings in other storage hardware costs of $81,000 annually, representing a total return on investment of 448 percent, and a payback period of 5.5 months. In addition to the financial value, the staff productivity increase was the equivalent of adding one full-time employee, and their average annual number of critical system failures was reduced from 17 to one.

Modernizing the IT infrastructure according to an Intelligent Abstraction model creates a foundation for optimized, simplified, more cost-effective operations. Times have changed. There is no more tolerance for segregated, disjointed architectures where each vendor, network, protocol, and location must be treated as a special snowflake. Instead, the modern storage environment enables unified, common data services across the entire infrastructure: information mobility, availability, automation, and operational efficiency.