The Economic Benefit of Autonomous Storage Operations with Tintri’s Intelligent Infrastructure

This ESG Economic Value Audit focused on the quantitative and qualitative benefits organizations can expect from leveraging Tintri’s workload intelligent infrastructure—specifically Tintri VMstore systems—as their optimized solution designed to move an organization toward autonomous operations and services delivery and away from manual processes. ESG created a modeled scenario that considered the operational benefits of deploying Tintri and the expected cost savings over a five-year period.

In a recent research survey, as shown in Figure 1, ESG asked organizations what their most important objectives were for their digital transformation initiatives. 56% of respondents selected becoming more operationally efficient as one of their most important objectives.¹ For most organizations, becoming more operationally efficient includes leveraging new technology such as AI, orchestration, data intelligence, and automation to improve decision making and drive down operational costs and overall IT spending. In the same ESG study, 49% of all respondents indicated that one of their digital transformation objectives is to adopt digital tools and automated processes, enabling users to interact and collaborate in new ways. As an organization begins to adopt smarter technologies, it drives innovation and frees up IT resources to focus more on business objectives and less on running day-to-day technical routines. Organizations focused transitioning toward automated processes become more operationally efficient, require fewer specialized IT experts, and can replace both the most mundane as well as challenging human-driven manual tasks with machine-learned intelligence that can ultimately lead to full automation.

When ESG asked respondents to describe the state of their IT digital transformation initiatives for 2021, half of the respondents indicated that they were currently implementing and executing on various digital transformation initiatives, but only 22% felt that they had mature systems in place. Another 16% were just beginning and 10% had not yet started any initiatives. While organizations may know what they want to achieve, results show that they are still in the early stages of digital transformation and may be looking for the right technology partners to assist them in achieving their goals.

Key features include:

• The right level of visibility and control: VMstore combines broad visibility across the infrastructure stack with granular performance and capacity metrics and visualization. Because there are no LUNs, volumes, or other traditional storage constructs to manage, the system provides immediate insight into latency, IOPs, throughput, and space usage at the managed object level—VM, database, or container level. This granular visibility and insight enable much greater efficiency because users can quickly respond to application needs and don’t have to spend an inordinate amount of time locating the source of bottlenecks and other problems. Granular cloning, backup, and restore are also tied to efficient operations. For example, it is faster and less disruptive to restore a single VM than an entire LUN.

• Self-optimization: This includes autonomous Quality of Service (Auto-QoS), which, among other things, ensures that each application incurs less than 1ms latency and will not experience any performance degradation regardless of system capacity utilization. Auto-QoS and other autonomous operations help VMstore customers avoid time-consuming performance management and costly resource overprovisioning. Because VMstore automatically and continuously adapts to application dynamics and changes to application load or specific user behavior, it also facilitates business agility. Organizations realize considerable enhancements to economic value and productivity because users aren’t hobbled by slow performance.

• Intelligent predictive analytics: VMstore leverages machine learning and three years of the customer’s historical data as a means of reducing capital expenses and the inordinate amount of time spent on cumbersome resource planning. VMstore can accurately predict future capacity and performance requirements up to 18 months in advance and utilizes sophisticated “what-if” analysis and scenarios to calculate the performance and capacity impact of additional applications over time to avoid over- and under-provisioning resources.

• Empowerment of end-users and IT generalists: VMstore is so easy to deploy and use that organizations don't need a team of storage experts and associated overhead, training, delays, trouble tickets, etc. Users are empowered by the architectural simplicity, which results in far fewer resources to monitor and manage compared to LUNs and volumes. This lower complexity also means that there are far fewer things that can potentially go wrong. Generalist-friendly administration improves agility and flexible resource management because any IT staff member can manage the storage environment. Developers and DBAs, for example, can easily manage the system and setup or tear down dev/test environments within minutes using zero-impact cloning that doesn’t affect production environments.

Figure 2. Tintri VMstore's Intelligent Infrastructure

The time-consuming tasks that are demanded of a traditional storage administrator include managing the storage environment and infrastructure, supporting application requirements, optimizing performance, protecting data, monitoring and reporting, capacity planning, evaluating features and technologies, and installing new instances. Over the years, many of these tasks have seen improvements in user simplicity (UIs), task automation (scripts and wizards), integration with other components (storage functions initiated through hypervisors or applications), and even added intelligence (guidance and alerts as a result of system analysis). Many storage vendors claim to have various levels of automation and intelligence built into their traditional storage systems but are not at the advanced level of Tintri.

Traditional storage solution vendors have attempted to build various types of intelligence and automation into their systems across multiple categories, starting with simplifying planned actions (through increased insight), simplifying the day-to-day operations (through assisted optimization), and, finally, attempting to simplify unplanned actions (through problem prediction and avoidance). Each of these sets of tasks becomes increasingly complex and difficult to deliver in an autonomous manner (especially as the environment grows and resources are added), and the limitations of traditional storage system code compatibility and infrastructure often prevent these capabilities from being effectively implemented.

The time-consuming tasks that are demanded of a traditional storage administrator include managing the storage environment and infrastructure, supporting application requirements, optimizing performance, protecting data, monitoring and reporting, capacity planning, evaluating features and technologies, and installing new instances. Over the years, many of these tasks have seen improvements in user simplicity (UIs), task automation (scripts and wizards), integration with other components (storage functions initiated through hypervisors or applications), and even added intelligence (guidance and alerts as a result of system analysis). Many storage vendors claim to have various levels of automation and intelligence built into their traditional storage systems but are not at the advanced level of Tintri.

Much like the automobile industry, which has incorporated features and systems to aid the driver, with the eventual goal of realizing a truly autonomous self-driving car, the storage industry similarly strives for fully autonomous storage systems that can be operated without a dedicated administrator. Autonomous storage removes siloed workflow bottlenecks (between storage admins, virtualization admins, DBAs, etc.) and empowers IT generalists to streamline delivery of services to the business. While no storage vendor is fully autonomous yet, Tintri has been leading this effort and working toward this goal for years, focusing on purpose-built Intelligent Infrastructure that avoids the limitations of traditional architecture and provides what are widely considered the highest levels of storage autonomy available in the market today.

ESG’s Economic Validation process is a proven method for understanding, validating, quantifying, and modeling the economic value propositions of a product or solution. The process leverages ESG’s core competencies in market and industry analysis, forward-looking research, and technical/economic validation. ESG performed technical validation, reviewed vendor case studies/economic models, and conducted in-depth interviews with end-users to better understand and quantify the impact of Tintri’s technology on their organizations, particularly in comparison with previously deployed alternative solutions. The findings were used as the basis for an economic model comparing the expected costs of operations with standard infrastructure to the costs after deploying Tintri VMstore.

With Tintri, organizations are set on a path to automate nearly all their storage-related tasks and create an environment where storage systems can respond more intelligently to the changing demands of modern applications and businesses. As organizations moved toward autonomous storage with Tintri, they found strong benefits in these three core areas:

• Autonomous Administration – Slashes operating costs and empowers IT generalists to deliver premium storage services for the modern organization. Autonomous operations reduce the skills gap while enabling operational freedom for end-users, IT generalists, and existing storage administrators.

• Faster Time to Value – Improves operating efficiencies, which accelerates time to value and delivers high infrastructure ROI at scale for enterprise workloads.

• Enhanced User Experience – Each stakeholder gains an elevated experience from benefits that meet persona-specific requirements, greatly improving satisfaction for end-users, freedom for IT admins, agility for IT managers, and predictability executives.

Modern businesses place great burden on the IT organization to provide agile cloud-like services to support and scale application demands that are constantly changing. Lack of coordinated visibility into storage, compute, and network layers with isolated management of virtual machines and databases limits business agility, making it difficult to provision, tune, or scale to meet changing demands. Tintri VMstore can be deployed, provisioned, scaled, protected, and recovered in a fraction of the time required by traditional systems, resulting in accelerated development cycles, faster time to value for application services, and reduced infrastructure risk.

ESG also validated that Tintri’s workload intelligent infrastructure can help reduce CapEx by avoiding the cost of overprovisioned resources and optimizing capacity for snaps, clones, and replicated data. Tintri also provides consistent support and maintenance costs over five years, eliminating the sharp increase in years 4 and 5 charged by most storage vendors. These savings, coupled with the substantial efficiencies provided via autonomous operations, result in a quick and substantively better return on investment (ROI).

• The Right Visibility, Control, and Outcomes for All – Tintri VMstore provides visibility and actionable insights for multiple personas from a single screen and allows for granular control, telemetry, and management from an application, VM, DB, or container level that is not possible through legacy storage-centric architectures, constructs, or administration.

• Satisfaction for End-users – Application admins, developers, engineers, and data scientists benefit from self-service capabilities, predictable and guaranteed performance, and sub-millisecond latency enabled by autonomous workload isolation and auto-QoS.

• Freedom for IT Administrators – Autonomous operations, supported with intelligence and intuitive actions, empower IT generalists, virtualization admins, DBAs, and developers to be more productive, responsive, and capable across the entire technology stack without the need for additional training or storage expertise. IT staff are able to complete tasks more efficiently without introducing workflow delays that require involvement of additional team members.

• Agility for the IT Manager – IT managers and directors are able to achieve their goals in a more agile manner by getting more from their junior administrators and less-experienced staff and enabling more flexibility around storage management staffing. Managers can gain greater insight into the impact of all production and test/dev behaviors on storage usage and reduce the number of helpdesk tickets substantially, while experienced staff members work on higher value and forward-looking projects. Tintri VMstore sets up managers to get more out of their limited resources and focus on proactive rather than reactive management.

• Predictability for the IT Executive – The detailed telemetry data and predictive analytics provided by VMstore enable IT executives to understand and predict system resources from a longer-term and more holistic application- or workload-centric point of view rather than simply from a storage-centric view. Executives can trust the Tintri machine learning models to accurately predict resource needs up to 18 months in advance. In addition to capacity requirements, they can proactively predict performance requirements and optimize resources, avoiding overprovisioning costs and improving efficiencies to get the most out of IT assets and personnel. This makes it far easier for the executive to achieve planning, budgeting, service level goals, and positive outcomes for the business.

Taking what we learned from our validation, ESG created a matrix that identifies a number of important storage attributes and application-enabling functions (tasks) and evaluates how easily and autonomously they can be performed across different categories of storage technologies. Figure 4 summarizes our analysis in a graphical manner. As shown by the columns, going from left to right, the more sophisticated the technology the less effort is required:

• Traditional or legacy storage systems that require tasks to be planned and performed manually.

• Traditional storage systems that can automate tasks.

• Intelligent storage infrastructure, which leverages autonomous operations, machine learning, and artificial intelligence.

• A forward-looking category where all tasks are performed by storage in an autonomous fashion.

While no storage system on the market today is fully autonomous, Tintri VMstore provides a higher level of autonomous operations than those that are limited by traditional storage architectures. Figure 4 summarizes our analysis in a graphical manner.

ESG assumed the organization currently has two full-time administrators managing services on a traditional storage system with 1PB of effective storage capacity. We then created a detailed storage administration model that considered the time required to manage typical storage-related functions over a five-year period. ESG then used the validated analysis summarized in Figure 4 to estimate and model the expected reduction in person-hours for each management. These calculations were based on the value of built-in intelligence, automation, and autonomous functions that could be expected from a leading traditional storage array with autonomous features built in and then with Tintri Intelligent Infrastructure. The results (shown in Figure 5) predict that Tintri VMstore can provide up to 89% lower cost of storage administration when compared with a traditional storage array. While autonomous features built on top of the traditional storage array provided significant administrative savings, the advanced autonomous features enabled by, and built into, Tintri’s Intelligent Infrastructure would still be expected to provide up to 78% lower cost of administration.

Next, ESG’s models estimated the expected total cost of ownership (TCO) over this 5-year period for the three storage arrays (assuming parity at a $/usable GB). ESG found that in addition to the administrative savings, Tintri VMstore can provide CapEx savings by avoiding the need to overprovision unused capacity (best practices for many traditional storage solutions suggest purchasing extra capacity to provide space for data services, garbage collection, and optimal performance while Tintri customers report running at 97% capacity with no performance degradation) and also by providing predictable, economical support and maintenance over a 5-year period. In contrast, many traditional storage vendors substantially increase the cost of support and maintenance contracts after the initial 3-year coverage period expires. ESG’s models predicted that Tintri can reduce the expected 5-year total cost of ownership by over a third when compared to traditional storage solutions (results shown in Figure 6).