Virtualization Technology News and Information
VMblog's Expert Interviews: SIOS Talks SIOS iQ and Machine Learning-based IT Analytics


As IT teams move more business-critical applications into virtual environments, they are becoming increasingly overwhelmed with reactive firefighting of performance issues.  They are still working and problem-solving in operational IT silos (network, computer, storage, and application) and using threshold-based tools to measure specific qualities (i.e. CPU utilization) that may indicate a problem.  This fragmented approach is outdated in today's complex, dynamic virtual environments.  

To dig in deeper and find out more, I spoke with Jerry Melnick, president and CEO of SIOS Technology Corp.  The company recently announced the newest version of its SIOS iQ IT analytics platform, which harnesses the power of machine learning and deep learning analytics to optimize the performance and efficiency of VMware environments. 

VMblog:  Tell me a bit about the latest updates to the SIOS iQ platform?

Jerry Melnick:  SIOS iQ is a new machine-learning based analytics solution that provides the precise, accurate information that IT needs to run business-critical applications in virtual environments without wasting, time, money or IT resources. It allows IT to move from reactive problem solvers to proactive strategists, implementing IT initiatives that add value to the core business functions. Our latest updates include the addition of performance forecasting - enabling IT to anticipate and eliminate performance issues before they materialize, and a new flexible, API-driven integration architecture that enables SIOS iQ to integrate data from a range of sources, including application monitoring tools and data aggregation fabrics.

VMblog:  How do existing approaches to identifying root causes of application performance problems fall short?

Melnick:  Optimizing application performance in VMware environments is incredibly complex, and the dynamic nature of these environments require highly advanced tools to automatically identify and recommend solutions to them. The current strategy of relying on multiple tools and teams to evaluate each IT silo: application, network, storage, and compute, leaves IT with a fragmented view of what's actually causing performance problems. With this approach, IT has to sift through thousands of alerts, compare analyses from multiple tools, and draw on their own knowledge and experience to optimize and solve problems. Alert fatigue quickly sets in with current tools that make it difficult to pinpoint which alerts are meaningless and which are worth diagnosing to solve a potential application performance issue.

Worse still, they have no way to predict potential problems, forecast performance or capacity requirements, or develop strategies to avoid problems from developing in the first place.  Without reliable, accurate information about their infrastructure, IT teams are forced to be reactive - fixing immediate issues as they happen and over-provisioning performance and capacity to avoid unexpected problems. In addition, few have the time or focus to think about the important role IT plays in supporting the company's core business objectives, such as staying agile, cost-effective, and responsive to customer needs.

VMblog:  How does the updated SIOS iQ improve on this approach?

Melnick:  SIOS iQ is a machine learning-based solution that learns patterns of behavior to help IT understand, optimize and solve application performance issues in virtual environments. This technique begins by using patented technology called topological behavior analysis (TBA) to automatically identify both the topology and the patterns of behavior among all interrelated components (VMs, network, storage, hosts) in the infrastructure. It continuously analyzes real-time machine data from sources across the infrastructure silos to understand the patterns of behavior between workloads and the resources they consume. It tracks the normal patterns of these complex behaviors as they change through work days, work weeks, and over the course of seasonal fluctuations.

This approach enables SIOIS iQ to precisely identify abnormal behavior in the context of the specific company's typical business cycle. SIOS iQ notifies IT only when actual problems exist and provides comprehensive information about the root cause, affected components, and recommendations to fix them. In this way, it eliminates alert storms typical of legacy monitoring tools and provides highly precise, accurate information about the virtual environment, root causes of problems, and the most efficient way to fix them. As a result, SIOS iQ frees IT from the daily grind of reactive problem handling to proactively operate and innovate in order to add value to their core business operations.

VMblog:  Other new features in SIOS iQ are intended to increase visibility across the IT infrastructure.  Why is this this important?

Melnick:  IT teams need smart tools that leverage advanced machine learning analytics to provide an aggregated, analyzed view of their entire infrastructure. IT teams using threshold-based tools have to review dashboards filled with charts, graphs, and data summaries. While some of these tools may present a map of the infrastructure that shows the hierarchy of physical relationships between components, none give IT the information they need to see how the components interact and impact one another. They leave IT to draw their own conclusions about how to solve problems and optimize their virtual environments.

In contrast, SIOS iQ meta-graph provides a powerful visualization of its deep learning analysis of thousands of related incidents. It shows the relationships and complex interactions between components and highlights the problematic interactions that are at the root cause of performance issues. The meta-graph reduces thousands of incidents identified as problematic to a small, manageable number of root cause problems. It provides clear insight into the components affected by these problems, and recommends specific steps to fix the problems.

VMblog:  What changed that created a need for machine learning-based IT analytics?

Melnick:  These days, organizations are moving more business critical applications to virtual environments that require consistent high performance. However, despite virtual IT environments being more complex than traditional on-premises environments, companies are still using an on-premises "siloed" approach to manage them. To make matters even more complicated, the IT infrastructure and the teams that manage them are siloed with each group (network, application, compute and storage) using a different collection of monitoring and diagnostic tools. This makes it difficult to find the root of application performance issues. As a result, IT is overwhelmed with time-consuming problem solving and many feel that they've becoming reactive firefighters solving the same problem repeatedly, wasting their valuable time, unable to use IT resources to their full potential.

VMblog:  And what does a machine learning-based IT analytics platform look like?

Melnick:  When end users report an application performance issue, IT teams from each of the different silos look at problem from their own narrow perspective, using a multiplicity of diagnostic tools. Some tools make primitive attempts at providing a broader view of the infrastructure by providing "dashboards" filled with multiple charts. This means IT is left to rely on crude comparisons of time stamps and charts, and their own subjective understanding of the situation to determine the likely cause. Finding the root cause of an issue in this way is extremely labor intensive, highly dependent on human judgement; and in most cases, requires rework later on.

Machine learning technology enables IT to analyze data from a wide variety of sources - across silos, and to account for the complex patterns of behavior between interrelated objects over time. Tools that use advanced machine learning and deep learning technology instantly identify the root cause of performance issues and provide recommendations for solving them with a level of precision and accuracy that humans alone cannot provide.  Machine learning based tools eliminate a lot of work, including the need to manually configure and continuously update monitoring tools.

VMblog:  How do you see the industry changing over the next few years?

Melnick:  The exponential growth that we are seeing in the size and complexity of virtual data centers has pushed IT departments past the limits of traditional, manual approaches. To get their arms around managing these large, complex data centers, more and more companies will be forced to implement automated data science approaches. The use of machine learning and deep learning technologies to understand and manage these environments will become the norm. Automation of data center operations will be driven by machine learning to allow dynamic response to changing requirements.  With virtual data centers that are automatically optimized by machine learning based analytics platforms, IT departments can turn their focus to adding value to their core business operations and end user productivity.


Published Friday, June 09, 2017 8:11 AM by David Marshall
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<June 2017>