Predictive analytics is delivering a lot of value when applied to enterprise storage. Tintri recently asked DCIG to write about their storage analytics, and I am glad they did. After taking a closer look, I believe that Tintri takes a fundamentally better approach to applying analytics to enterprise storage than many other vendors. Most other vendors start with break-fix support. Tintri analytics start with optimizing application performance.
I remember DCIG’s first encounter with Tintri. Our founder, who had introduced virtualization at a large financial services firm before starting DCIG, was amazed by two things he observed at a Tintri client’s site:
1) the large number of virtual machines served per Tintri VMstore appliance, and
2) the tiny effort required to manage the storage.
What was true then about running hundreds of VMs on a small hybrid storage array is still true today when running thousands of VMs on a Tintri all-NVMe array.
The key is the way in which Tintri built analytics into the foundation of the VMstore design.
Tintri’s integrated real-time analytics actively and automatically optimize application performance from day one. Beyond day one, Tintri’s predictive analytics enable ongoing workload optimization for even the largest virtual machine infrastructures.
Tintri analytics are native to every aspect of managing VMstore resources. As a result, VMstore all-NVMe appliances can run thousands of virtual machines and SQL Server databases (and soon container volumes) with astonishingly little storage management effort.
Results that matter to Tintri customers are:
- high levels of application availability
- consistent performance across time
- maximum infrastructure agility
- minimum IT infrastructure and management overhead
Predictive Analytics Fundamentals
Telemetry Data is the Foundation for Predictive Analytics
Many enterprise storage arrays transmit not just fault data but extensive additional telemetry data about workloads back to the vendors. This data includes IOPS, bandwidth, and latency associated with workloads, front-end ports, storage pools, and more.
Most vendors evaluate the collected fault data and advise customers how to resolve problems, or they remotely log in and resolve problems for their customers. Some vendors apply predictive analytics and machine learning algorithms to data collected across the entire installed base to identify potential problems and optimization opportunities for each array in the installed base.
Visionaries Pursue Autonomous Infrastructure Optimization
Most enterprise storage vendors now offer some level of Artificial Intelligence for IT Operations (AIOPs). Most vendors approach AIOps from a support perspective that focuses on reducing application downtime by exposing component failures and helping troubleshoot performance problems. Some vendors have added predictive and cross-stack analytics to their support tools to reduce application downtime further. These advances can have a meaningful impact on availability and data center operations.
However, the end goal of predictive analytics for the more visionary providers goes beyond eliminating downtime. Their goal is to enable data center infrastructures to autonomously optimize themselves for application availability, performance, and total cost of ownership based on the customer’s priorities.
Vendors who commit to this path and execute better than their competitors create real value for their customers. The benefits include:
- Measurably reducing downtime
- Avoiding preventable downtime
- Optimizing application performance
- Significantly reducing operational expenses
Predictive Analytics Features That Matter
Proactive interventions and recommending configuration changes are two predictive analytics features that contribute the most to reducing downtime and optimizing application performance.
Proactive interventions identify something that will create a problem and then notify clients about the issue. Interventions may include providing guidance on how to avoid the problem. A wide range of interventions is possible, including identifying when an array will reach full capacity or identifying a network configuration that could create a loop condition.
Recommending configuration changes enhances application performance at a site by comparing the performance of the same application at similar sites, discovering optimal configurations, and recommending configuration changes at each site.
Tintri Actualizes the Predictive Analytics Vision
Tintri Analytics Focus on Optimizing Application Performance
Tintri focuses its analytics on optimizing application performance, not break/fix support. This focus is a fundamentally better approach to applying analytics to enterprise storage than many other vendors offer. As a result, Tintri delivers all four of the benefits mentioned above to a degree that is astonishing to experienced virtualization professionals.
Plus, Tintri’s proactive interventions go beyond notifying clients about problems and providing guidance on how to resolve those problems. Tintri’s integrated VMstore analytics actively avoid problems. Tintri also goes beyond merely recommending configuration changes. Instead, Tintri enables one-click workload optimization.
Optimizes Application Performance Automatically
Tintri’s TxOS storage software applies analytics to telemetry data in real time to operate the VMstore. TxOS knows the IO characteristics of each SQL Server database, virtual machine, and Tanzu persistent volume—Tintri calls these “managed objects.” It also knows which IO requests map to which managed objects. Thus, TxOS knows how much latency each managed object is experiencing and uses AI to dynamically adjust performance resources among even thousands of active workloads to optimize application performance, delivering consistent sub-millisecond responsiveness to each workload.
VM-aware Storage Enables Application-level Management
Second, VMstore—as its name implies–is designed for one of the most challenging storage environments, the highly virtualized enterprise data center. Such data centers may support tens of thousands of applications and may also serve up thousands of virtual desktops.
Because Tintri designed VMstore specifically for virtual infrastructures, it foregoes traditional volume and LUN storage constructs. Deep integration with VMware and other hypervisors enables VMstore to automate application performance management and data protection for each managed object. This automated management applies equally to virtual machines, SQL Server databases, and Tanzu persistent volumes; and is dramatically more efficient than managing volumes and LUNs.
The integrated analytics can use vSphere tags in conjunction with Tintri service groups. Thus, protection policies, snapshots, replication, and more can be based upon tagging in vSphere. This is important because communication failures between virtual server managers and storage administrators can result in incorrect workload placement and protection. Once vSphere tag-based workload placement and protection policies are in place, organizations can avoid these missteps. Enterprises can also use tags to group managed objects for reporting, trending, and planning.
Analytics that Make an Impact
Tintri’s integrated real-time analytics actively avoid problems from day one, enabling VMstore all-NVMe appliances to run thousands of virtual machines and SQL Server databases (and soon container volumes) with astonishingly little storage management time and effort.
All of this comes from Tintri having built analytics into the foundation of its VMstore product with a focus on actively and automatically optimizing application performance.
This is why DCIG says Tintri actualizes the predictive analytics vision. And that makes Tintri a predictive analytics visionary.
Keep up to Date with DCIG
To be notified of new DCIG articles, reports, and webinars, sign up for DCIG’s free weekly Newsletter.
To learn about DCIG’s forthcoming research and publications, see the DCIG Editorial Calendar.
Technology providers interested in licensing DCIG TOP 5 reports or having DCIG produce custom reports, please contact DCIG for more information.