Stop Collecting – Start Automating

The technology that powers VMTurbo’s Operations Manager product approaches managing IT environments in a fundamentally different way. While the results are important – application assurance, improved utilization of IT resources, and a reduction in the costs to run cloud and virtualized data centers – it’s the approach that makes it innovative. Rather than focusing on collecting, alerting and troubleshooting, the technology uses an economic approach to making and automating operational decisions in the data center.  By representing the IT environment as a market of buyers and sellers (applications and physical resources) and using supply and demand principles to establish relative prices, the solution enables “the invisible hand of the market” to determine resource allocation decisions that reduce contention and prevent problems from occurring.

Virtualized infrastructure introduces new challenges for managing IT environments and how to best allocate compute and storage resources to prevent workload interference and ensure application performance while getting the most out of your physical resources. Other technologies focus on collecting and monitoring vast amounts of data, establishing thresholds for individual metrics, and then generating alerts when utilization of servers, storage or virtual machines resources reaches dangerous levels.

While this is can be helpful in identifying infrastructure bottlenecks which are currently (or soon will be) impacting application performance, it does not provide the intelligence to make the right resource allocation decisions to resolve performance issues (or avoid these from happening in the first place).  Making these decisions is not a trivial task due to the number of variables which must be considered to prevent resource contention when virtualized workloads share common server and storage infrastructure.

Intelligent Workload Management and the Economic Scheduling Engine

Given the complexity and dynamic nature of virtualized infrastructure, it is not possible for even the most experienced human resources to solve this problem in a scalable fashion. A sophisticated, mathematical (analytic) approach is required to drive this decision making. We refer to this approach as Intelligent Workload Management – providing a complete closed loop of VM monitoring, analyzing and automating to simplify the process of identifying & resolving performance bottlenecks.

VMTurbo drives intelligent workload management with an Economic Scheduling Engine that continuously tunes the environment to reduce resource contention and avoid performance issues before problems occur. The technology represents the virtualized IT environment as a marketplace. It conceptualizes the infrastructure as a service supply chain and balances workloads according to supply and demand – based on price and budget priority. At any time, there is a certain supply of CPU, memory, drive space, IOPS, and other resources demanded by virtual machines and applications. Resource prices fluctuate according to utilization rates as demand changes. In this abstraction layer – as in any market –buyers are constantly shopping around for better prices and sellers raise price as resources become scarce.

High priority, business-critical applications are given more currency to spend on resources. Should contention in the environment become elevated, resources become over-utilized or scheduling conflicts occur, those applications are able to afford the resources needed to preserve performance and reliability levels. Lower priority applications utilize lower priced services. Conflicts are automatically resolved by the market according to business requirements.

The Economic Scheduling Engine uses a mathematical approach based on economic principles of supply and demand to derive a broad set of resource allocation decisions such as workload placement, workload sizing, storage resourcing, application sizing, and overall capacity adjustments. The technology makes these decisions by analyzing real time and historical workload demand, and determining how it can best be accommodated through the available server and storage capacity, or by provisioning or consolidating capacity.

This decision making process takes into account a broad set of performance and configuration constraints which naturally restrict where in the infrastructure workloads can run. In addition, the engine ingests pre-configured affinity and anti-affinity rules and provides the ability to configure additional constraints to restrict where specific workloads can be placed to meet business goals such as business continuity.

These resource allocation decisions (recommendations) can be fully executed by the system (manually or automatically). VMTurbo Operations Manager can also be integrated with third party change control systems so that these actions can be approved and scheduled as part of the organizational change management process.

Real-time decision-making no new databases

VMTurbo Operations Manager does not rely on Virtual Center/vSphere or other hypervisor databases. The technology collects metrics through the native APIs in the virtualization platform (Virtual Center SDK API, XenAPI) which provide access to 20 second samples of data for all critical performance metrics. VMTurbo uses the peak and average data points in its analytics, which run in memory, to enable fast real time decision making. This is critical for operations staff that typically has to troubleshoot and resolve performance problems as quickly as possible, in response to complaints from users about poor application performance.

VMTurbo Operations Manager also includes extensive reporting capabilities through a large and ever expanding library of standard reports. It also has a custom reporting capability which enables users of the system to build their own reports from any of the metrics/data collected by VMTurbo.

Included in the standard reports are specific reports which highlight dormant virtual machines, wasted storage and associated files that can be deleted, virtual machines which can be re-sized including recommended configurations. Because VMTurbo views monitoring, alerting and reporting as a commodity, we offer this functionality for free in its Virtual Health Monitor product.


 

Free Trial & Demo CTA

Free Trial request A demo

White Paper Promo: Applying an Economic Model to IT Management

Applying an Economic Model to IT Management

Our Unique Top-Down Approach

Our Unique Top-Down Approach