With an often diverse mixture of legacy platforms and the need for an increasing number of storage systems brought in to support more – and ever more virtualized – environments, many organizations are facing an increasingly problematic state of “storage sprawl.” This sometimes chronic condition results in a scarcity of valuable data center floor space, an infrastructure imbued with multiple, sometimes widely divergent management and administration workflows, as well as disappointingly low levels of storage utilization and unpredictably spiraling costs.
These same organizations are also invariably faced with a constant sense of urgency to meet their business objectives and remain agile enough to cope with the introduction of new ways to build and sustain competitive advantage and mounting workload volumes. To effectively address these challenges, IT leadership must plan more strategically, ensuring they take action early to avoid, and remedy, the effects of storage sprawl when it comes to IT planning and procurement processes.
A solution that many firms are already embracing is systematically consolidating their disparate storage systems onto a single, more efficient platform capable of adapting to the widest possible variety of workloads. Companies that have implemented this have achieved multiple benefits including increased utilization, lower OpEx, and more efficient and repeatable management and administration workflows.
While this solution sounds like the perfect fix to the storage sprawl challenge, there are some key focus areas to consider before heading down this road. Here are seven important characteristics to look for in any viable storage consolidation platform:
Within a consolidated platform, there are no longer highly specialized solutions to accommodate individual workload environments. Therefore, the ability of the new platform to easily adapt and deliver optimum, sustained performance across a number of intermingled, dynamic and varying workloads is vital. The system must be capable of optimally balancing the resource allocation across many workloads with minimum to no human intervention required. A good consolidation platform will have the ability to autonomically alter its resource allocation policies and behavioral characteristics in real time to accommodate varying and often unpredictable fluctuations in workload demand. An intelligent, self-tuning storage platform will mitigate the need for storage administrators to invest their valuable time in the constant iterative cycle of measuring, tuning and validating performance across a broad and deep workload mix.
Consolidation of workloads onto a single platform means that reliability requirements are dictated by the most critical workload(s), which sets the bar for all others. Component and systems reliability become of paramount importance as the fault domain essentially decreases. This concept of the “blast radius” of any consolidated storage platform mandates that overall system reliability must be sustained at the highest possible level at all times to ensure that all workloads are always optimized and available.
The ideal consolidation platform will support increased storage utilization while removing the need to spend on costly storage services and excess capacity. Data centers with substantial storage sprawl have a tendency to experience fairly low utilization levels because there is a need to service many workloads on an individual, server by server – sometimes workload by workload – basis. Here, storage systems are often not being used to their fullest capacity as these pockets or islands of excess capacity proliferate. This is both wasteful and drives up acquisition and operating costs. The goal is to aim for greater/higher utilization on a smaller number of systems, thus increasing operational efficiency and leveraging economies of scale. The multiplicative effect of over-provisioning (and often over-investing) in excess capacity across many disparate storage pools to accommodate future growth is mitigated when multiple pools can be consolidated into a single, efficient environment.
As data center space is increasingly at a premium, a consolidation platform with higher areal density (the volume of data that can be stored in gigabytes per square inch), will be more cost-effective. Consolidation will increase data center capacity and cost efficiency, as there are fewer media devices capable of storing more data and providing more favorable power consumption ratios. So the best outcome can be achieved by aiming for the platform that can provide the right balance of a) the highest volume of data that can be stored per square foot of data center space, and b) a low power consumption ratio.
5. Ease and simplicity
Staff productivity often deteriorates as multiple disparate storage platforms proliferate throughout the data center. Admin staff are required to develop skills and remain current on multiple management platforms and tools. With the introduction of a single, consolidated management platform – especially one that is likely to be more streamlined and potentially easier to learn and use – any training for the new platform should be quick and efficient, and thus reduce the time between deployment and operational go-live. In addition, without the quagmire of systems causing storage sprawl, much of the complexity in management can be mitigated by simply reducing the number of managed endpoints.
The consolidated single platform must be capable of handling large and growing volumes of data and significant working set sizes. Working set sizes will vary depending upon the application and workload environment. As such, the platform must be able to handle workloads with potential scope of between less than a terabyte to multiple petabytes. Future requirements must also be taken into account in terms of performance and capacity, as the workload mix is certain to change over time. So scaling for each of those aspects separately will be important to ensure the environment is designed to suit evolving workloads.
While the all-flash data center is often discussed with much fanfare, in reality all-flash arrays are not well-suited for consolidating workloads given the undue strain they put on IT budgets and their reliability and scale limitations. That’s not to say flash doesn’t have an important role to play in a storage consolidation platform, however. When hybrid arrays make use of a cost-effective and intelligently architected combination of SSD and HDD – a flash-optimized approach – they are well positioned to deliver the performance and reliability required for consolidating the majority of workloads onto a single platform at a predictable, sustainable price point.
Storage consolidation can represent a great way to both reduce cost and enhance operational efficiency and IT organizational agility. A properly designed and implemented storage consolidation platform can lead to a number of highly sustainable benefits in any diverse, growing organization. However, careful planning and solution vetting, as well as rational consolidation best practices, are required to ensure that these benefits are realized to the fullest extent.
Randy Arseneau is CMO at storage company Infinidat.