Turbonomic Application Resource Manager (ARM) is a fantastic AIops platform to ensure that performance is assured from Application through to Array while maintaining operational governance and efficiency. In this article we focus on 5 ways to get the most out of Turbonomic ARM when it comes to Storage Operations with a real-world example at the end.
1. Give applications the storage performance they demand without touching a keyboard
Turbonomic is uniquely positioned to observe everything from the application to the array. When applications demand more storage performance regardless of whether they are containerized or virtualized, Turbonomic observes this demand through the entire stack.
At the same time, Turbonomic is constantly observing the storage infrastructure supply. Matching demand to supply in real time, Turbonomic generates actions to migrate application workloads to the best storage to deliver the best performance.
These actions are governed by operational policy and can be automated so that applications always get the right storage performance when they need them.
2. Know what applications are using your storage
Turbonomic is observing the entire stack from Application to Array and everything in-between. Whether you start at the Storage Controller node, the array or logical pool / volume and work upwards you can immediately see what applications, containers or VMs are consuming from that starting point and vice-versa.
So, if you have disruptive maintenance to perform on any storage component you can quickly identify the virtualization and application components using that component and ensure that any change approvals are targeting the right teams and business owners.
3. Show that storage congestion was not the cause of application performance degradation
If application performance degradation occurs it’s important to quickly identify any resourcing issues that might be contributing to the incident. As a Storage Operations team it is vital to quickly isolate what storage is being used by a particular application suffering performance.
Turbonomic lets you see the entire supply chain underneath a given application and see immediately whether storage infrastructure is a risk contributor to the application performance issue or not.
Furthermore, if there is performance risk within the storage infrastructure, Turbonomic may have already dealt with it automatically by migrating workloads or provisioning additional capacity or will show the actions necessary to restore optimal performance.
4. Run simulations to estimate storage requirements for workload migrations
Businesses need applications spun up quickly and need more storage as demand grows. Turbonomic creates abstract digital twins of your environment allowing you to apply it’s AI based analysis offline for different scenarios to answer the important questions about whether there is sufficient storage to accommodate new application workloads, or what additional storage is needed to migrate workloads from one storage system to another. Different scenarios can be played out using the abstract digital twins of actual application workloads and storage.
5. Automatically scale and know when to provision storage to meet demand
When new unexpected demand arises, Turbonomic is constantly making AI based decisions about how best to place and size resources to meet the demand. These decisions result in tangible actions many of which can be completely automated so that when demand surges you can rely on Turbonomic to scale and provision new storage to meet that demand within the constraints of your operational policies.
What does this look like in practice?
A business needs to respond to an unexpected market event that is driving customer interactions through the roof. Customer call center agents are experiencing delays on the back end systems resulting in long call wait times and a degrading customer experience. This increased activity is translating into a surge in IO operations for some applications which is causing a rise in latency on legacy storage infrastructure based on older Flash technologies. Some of these applications are containerized on OpenShift using Virtual Machine nodes, others are running natively inside Virtual Machines on VMWare vSphere. IBM FlashSystem volumes are available to the vSphere hosts. Using Instana Application Performance Management it is clear that response time service levels are close to being breached for some of the critical call center applications.
How would Turbonomic manage this?
Unlike human operations teams, Turbonomic observes your infrastructure from application to array all the time, making decisions about resourcing that result in actions that can be automatically executed. In this scenario, Turbonomic understands the supply chain relationship all the way from the impacted critical business applications being observed by Instana down through OpenShift containers and vSphere Virtual Machines all the way to the vSphere hosts and legacy storage devices underneath. When applications demand more IOPS this translates through to increased IO on the VMs supporting the application. One advantage of virtualization is the ability to non-disruptively migrate workloads across physical infrastructure. Integrations with Instana, OpenShift, vSphere, legacy storage and IBM FlashSystems allow Turbonomic to observe the entire stack in real time. Turbonomic decides that some vSphere VMs need to migrate from the suffering legacy storage with increasing latency to the newer IBM FlashSystem storage where there is sufficient IOPS capacity to meet the surge in demand and restore response times to acceptable service levels. Turbonomic automatically instructs vSphere vCenter to migrate VMs from the legacy storage devices to the IBM FlashSystem volumes to ensure that application demands are satisfied and storage congestion does not become a critical factor in application performance. Furthermore, if total demand nears the available supply capacity, Turbonomic will automatically generate actions to expand storage volumes ahead of time or move less demanding workloads off onto other storage so that there is always enough storage available. Applications get the resources they need when they need them automatically. Once the surge in demand is over, Turbonomic will recommend that storage be released for use elsewhere to maintain efficiency goals.