How to Get Smarter About Availability
By choosing a high-availability solution that is smart, your enterprise will enjoy the flexibility it needs to adapt and thrive in today's business climate.
- By Don Boxley
- November 28, 2018
Almost all enterprises today require the highest levels of availability for their data and applications. Even the shortest amount of downtime can lead to both immediate and long-term business loss (not to mention a potential hit to one's career).
IT professionals are now realizing that even the newest high-availability (HA) technologies (hardware and software) and methodologies are unable to truly support today's dynamic business requirements. The drain on budget and man-hours spent testing and retesting to ensure reliable performance is becoming intolerable; the time would be better spent on activities that are more directly tied to bottom-line success.
What's needed is not higher availability but "smart" availability. In other words, what's required is what is being demanded across every other area of today's enterprises -- the ability to build intelligence on top of basic principles. In this case, it means layering intelligence on top of the basic principles of high availability.
Traditional HA Drawbacks
Most IT professionals will agree that high availability can be simply defined as the continuous operation of systems and applications. Most will also likely agree that although HA can be accomplished in a variety of ways, each method is typically associated with drawbacks from an IT resources, manpower, and/or cost standpoint.
One common example relates to failovers, in which system components are moved to a secondary system for scheduled downtime or unplanned failures. Clustering methodologies are frequently utilized with this approach to make resources between systems -- including servers, databases, processors, etc. -- available to each other. Clustering is applicable to virtual machines (VMs) and physical servers and can assist in enabling resilience for operating system, host, and guest failures. Failovers involve a level of redundancy that necessitates sustaining high availability by involving backups of system components. Redundant storage and networking options can be leveraged with VMs to encompass data copies and/or system components.
For some applications, high availability is not necessary. Yet, if an application sits on the same resource as high-availability apps, then it is likely receiving the same level of dedicated attention and resources. For instance, some of the more comprehensive hypervisor-based platforms are rather indiscriminate in this regard. Therefore, users may end up paying for high availability for components that don't need them.
Also, traditional high-availability approaches involve constant testing that can drain human and financial resources. Even worse, neglecting this duty can result in unplanned downtime. Arbitrarily implementing redundancy for system components broadens an organization's data landscapes, resulting in more copies and potential weaknesses for security and data governance.
Virtualization Methods Losing Relevance Due to Digital Transformation
Many virtualization methods for HA are losing applicability due to digital transformation. To really revolutionize the way your company does business with digitization technologies, you must implement them strategically. Historical HA approaches simply don't allow for the fine-grained flexibility required to optimize business value from digitization. Digital transformation means accounting for the varied computing environments of Windows and Linux operating systems as well as containers. It means integrating a variety of legacy systems with newer assets purposely designed to manage the flood stemming from big data, AI, analytics, and other modern systems.
Most important, digital transformation means aligning infrastructure with business goals in a way that supports continuously evolving domain or customer needs. Such flexibility is critical to optimizing IT processes around the goals of end users. The reality is most conventional methods of HA add to the infrastructural complexity of digital transformation, but do not address the primary requirement of adapting to evolving business requirements. In the wake of digital transformation, organizations must restructure their various IT systems around domain objectives, as opposed to structuring their objectives around existing system capabilities, which simply decreases efficiency while increasing costs.
Being Smart About Availability
Today, being smart about availability doesn't just mean ensuring the continuous availability of the system, application, or workload. It means moving the workload so its unique service levels can be met, as well as other considerations (including performance, security, and cost). In other words, to be smart about availability means moving it to its best execution venue -- i.e., moving data to where it can be best be served, and consequently serve you -- thereby positioning you for maximum competitive advantage.
Today, there are technologies and methodologies that enable you to be smart about your availability -- to move workloads independent of one another, which is one of the key disadvantages of traditional physical or virtualized approaches to workload management. By disengaging the array of system components (such as application workloads, services, share files, and containers) without standardizing on a single operating system or database, these technologies and methodologies can transfer the components to the environment that works best.
Smart Availability Is a Smart Judgment Call
Smart availability's "judgment call" is based on how best to achieve a specific business objective. These technologies and methodologies offer flexibility for individual instances to ensure virtually zero downtime and a seamless transition from one compute environment to the next.
The use cases for this near-instantaneous portability are countless. Companies can use these techniques for uninterrupted availability, integration with new or legacy systems, or the incorporation of additional data sources. Most of all, they can do so with the assurance that the intelligent routing of the underlying technologies is selecting the optimal setting to execute workloads. Once properly architected, the process takes no longer than a simple stop and start of a container or an application.
A smart availability solution offers all of the advantages of HA at a lower cost and with a much greater degree of efficacy. Furthermore, it delivers the agility required to capitalize on digital transformation, enabling organizations to move systems, advanced applications, and workloads such as big data, AI, and analytics to where they can best serve, drive competitive advantage, and contribute to the bottom line.
By choosing an availability solution that is smart over one that is simply high, your enterprise will enjoy the flexibility needed to adapt and thrive in today's business climate, where the only constant you can depend on is change.
Don Boxley Jr. is a DH2i cofounder and CEO. Prior to DH2i, Don held leadership roles at Hewlett-Packard where he was instrumental in sales and marketing strategies that resulted in significant revenue growth in the scale-out NAS business. Boxley spent more than 20 years in management positions for leading technology companies, including Hewlett-Packard, CoCreate Software, Iomega, TapeWorks Data Storage Systems, and Colorado Memory Systems. Boxley earned his MBA from the Johnson School of Management, Cornell University.