BEZVision Technology

BEZ Systems is an innovator in the emerging market for predictive analytics solutions. Predictive analytics provide a line-of-business view of application and database resource utilization that allows the enterprise to accurately profile both current and future application performance, compare change and growth alternatives, forecast results to set expectations and verify actual performance results versus predictions. BEZ Systems empowers IT to manage change, set realistic expectations, justify provisioning actions and ensure the delivery of consistent, uninterrupted data service to the business at the lowest cost.

For over 20 years, BEZ has been expanding and licensing BEZVision, with unique technologies that include:

Performance Modeling

Performance Modeling is at the core of the BEZVision technology platform and is driven by hourly profiles by collecting data 24 X 7 and creating two types of predictions, "As Is" predictions and "What If" predictions.

"As is" Predictions

"As-is" predictions are generated, and results are saved, automatically in BEZVision. This type of prediction uses historical data to calculate anticipated growth trends over the next 12 months, assuming that there will be no changes to the environment. The rate (daily, weekly (default), monthly) at which data is collected is set when you configure the model. When results are generated, they are compared to previously entered performance goals. These can become the basis of performance goal extensibility advice. As-Is predictions require a user-specified model and the number of periods to forecast (default is 12; period size default is one month). Growth forecasts are calculated based on historical data or standard growth plans by percentage per period or for the entire forecast. As Is predictions take historical trends of user activity and disk growth, and project those forward assuming no hardware or storage configuration changes.

"What if" Predictions

"What if" Predictions are manually initiated, based on data that you supply, including:

  • Configuration data – product hierarchy relationships; hardware configurations; OS; DBMS configurations
  • Measurement data (by Workload) – source period selection (representative time period to be used for modeling); resource utilization data (CPU, memory consumption); performance data (response time and throughput metrics)
  • Planning/Workload forecasting information – prediction category selection; prediction plans (input on what is changing); prediction timeline horizon setting (number of periods and period size)
  • Performance goals/service level objectives (SLO); user preferences (for performance goal setting)

In addition to greatly reducing the time and effort required to effectively measure and model a system, the BEZ technology holds the answers to the following important "what if" questions that cannot be answered satisfactorily with traditional performance management technology today:

  • What if I could see ongoing performance metrics by specific Lines of Business?
  • What if I had a way to understand the resources a planned new application is really going to take, how it will perform and how it will affect other applications after it is fully built, deployed and tuned?
  • What if I could visualize the near-term and long-term impacts of adding a new application or changing a database design – before implementation?
  • What if planning groups could be more proactive and had a methodical process that was easy to use, non-intrusive, continual, and repeatable?
  • What if I could accurately understand the impact of upgrades or migrations before deployment?
  • What if I had a framework to collect historical data and use statistical process control to determine which metrics indicate trends and which are merely temporary anomalies?
  • What if I could see out 6-12 months and estimate performance to reduce the time and cost of testing, minimize performance risks, justify necessary upgrades and build confidence in delivering consistent and acceptable service in the future?

BEZ advanced modeling technology satisfactorily addresses all of these "what if" questions. Direct answers to questions such as these help the enterprise:

  • Reclaim lost budget dollars
  • Reduce frustration
  • Free up personnel time
  • Steer clear of emergency budget demands for more hardware
  • Avoid the need for heroic tuning efforts to just get through another peak period

Workload Analytics

BEZVision delivers performance information as a series of workload analytics. A workload is simply a body of work that is meaningful to business and IT management. BEZVision transforms the performance data captured by multiple sources into actionable information that is easy for management to understand and to use to support critical IT decisions.

Beyond providing analytics with a business oriented understanding of resource utilization, performance, and service levels, BEZVision provides an automated early warning system that will identify and notify designated users of abnormal trends, performance objective breaches, and actual vs. expected variations for problems that may be just over the horizon or well into the future.

BEZVision workload analytics also provide a view of workload performance information in the past, present, and future. Over time, BEZVision allows IT management to migrate from near-term tactical responses to long-term strategic planning.

Detailed measurements, workload-oriented analytics and powerful predictive capabilities combine to help IT align more closely with the needs of the business. With BEZVision, IT can model and understand the impact of:

  • Activity growth
  • Database size or I/O processing growth
  • Hardware changes (CPUs, CPU speed/quantity, Number of Nodes, Vendor, Model, etc.)
  • Storage related changes (vendor, model, disk type, raid setting, number of disks, etc.)
  • Schema changes
  • Migration to cluster from standalone instance
  • Cluster expansion (additional nodes/instances into an existing cluster)
  • Consolidating workloads from different servers
  • Adding or removing workloads

Performance Repository

Raw data is collected from different sources within the infrastructure; the Operating System, Database and Disk sub-system. This data is aggregated and aligned on an hourly basis to provide a comprehensive profile of the data service delivery that was attained. BEZVision supports multiple data collection methods across multiple platforms. Please see the Supported Platforms Page for additional information.

End-to-end performance management has been a goal for years but attaining that goal has been rather elusive for various reasons: today's information systems are more complicated, more modular, and more challenging than ever to manage. Furthermore, end-to-end performance management has far more utility when applied to tuning efforts or immediate problem resolution, but is rather myopic when trying to understand and manage service delivery to the actual mix of business and other workloads. In order to effectively manage today's highly distributed application environment, IT managers need a single unified view of both the static elements of the infrastructure (e.g. a Configuration Management Database or CMDB) and of the dynamic interactions between and among the various components while applications and users are competing for service (e.g. a Performance Management Database or PMDB). When combined, IT management has a unified and consolidated view of application management with an eye toward maximizing service delivery to specific high-priority business workloads and optimizing application performance for all system-wide workloads.

What we do in 2 minutes
BEZVision is a predictive analytics solution for application and database environments. If you are tight on time, check out our 2 Minute Explainer.