|Empirical I/O Performance Metrics|
The I/O operation metrics collected by hIOmon are empirical metrics. That is, they reflect the tangible observations made by the hIOmon I/O Monitor software component. These observations are based upon I/O operation trace data captured at points within the operating system where and when the individual I/O operations are actually processed. The hIOmon software leverages the Microsoft® "Event Tracing for Windows (ETW)" facility for capturing this I/O operation trace data in an efficient and reliable manner, including in production environments.
Since they are empirically-based, the hIOmon I/O metrics reflect exactly how an application or workload actually performs its I/O operation activity. Moreover, these metrics can reveal how an application or workload actually experiences its I/O operation activity in terms of performance (together its "I/O Profile").
The empirical I/O operation and performance metrics uncovered by hIOmon enable you to compare your actual numbers to those mentioned within benchmarking test results/reviews, vendor performance claims, and within various "rules of thumb".
|Pertinent I/O Performance Metrics|
Empirical I/O metrics provide the important base. Empirical I/O metrics that are moreover pertinent to your particular applications, workloads, and devices are paramount towards addressing your specific interests and needs. These interests and needs can be related to determining storage requirements and purchases, performing performance tuning and problem diagnosis, and undertaking product comparisons.
hIOmon allows you to gather empirical metrics that are relevant to your very own applications, workloads, and devices upon a specific, individual basis. In addition, these metrics can be collected by hIOmon upon your own computer system during everyday usage.
Other specific use-cases include determining "Quality-of-Service (QoS)" requirements and furthermore confirming that these requirements are in fact being met.
Running within a virtualized environment (including within a cloud virtual machine)? Use hIOmon to independently verify actual I/O operation usage and performance.
|Automatically-Aggregated Summary I/O Metrics|
Summary metrics representing I/O operation activity are generally derived by collecting and then separately post-processing large (often vast) amounts of individual I/O operation trace data. This post-processing incurs saving/storing all of the trace data along with the time and effort required in performing the subsequent post-processing – all of which can be impractical, especially with long trace periods.
hIOmon features the ability to automatically accumulate summary I/O operation metrics at the time when the individual I/O operations are observed in realtime. This provides the accuracy and precision of an I/O operation trace regarding summary metrics but without the various resources, time, and efforts required to subsequently post-process I/O trace data.
Moreover, hIOmon enables users to maintain their privacy and security since there is no need to send any of their I/O operation trace or other data to hyperI/O for summarization or any other processing. (To additionally enhance security, hIOmon software components are digitally code-signed to ensure that the software came from hyperI/O and to protect the hIOmon software from subsequent alteration).
hIOmon can aggregate the summary metrics upon an individual specific file, device, or process basis. In addition, the summary metrics can be collected upon a user-specified periodic basis for offload to an export file. For instance, the WIOP configures the hIOmon software by default to collect cumulative summary I/O operation metrics that are offloaded/written to CSV-formatted export files upon a ten-minute periodic basis.
hIOmon fosters a "top-down" approach towards acquiring and analyzing I/O operation performance metrics. Rather than having to first or solely become immersed in volumes of I/O operation trace data details, the hIOmon summary metrics enable you to quickly see at the very outset key metrics related to I/O operation activity and performance – including I/O Profile specifics, even at a glance.
Going far beyond typical overall average and maximum value metrics, the hIOmon Top-Down approach features metrics such as the hIOmon "Performance Threshold Range Metrics". These innovative metrics enable you to immediately view I/O performance from the perspective of discrete ranges of I/O operations per second (IOPS), megabytes-per-second (MB/s) data transfer rates, response times (RT), I/O operation queue depths/lengths (QD), and data transfer sizes/lengths (DTS) actually observed by the hIOmon I/O Monitor for individual devices, files, and processes.
The hIOmon DXTI metric further accentuates the Top-Down approach. Just as "fuel economy" (i.e., "miles-per-gallon") is likely a major starting point in considering your automobile's performance, the DXTI metric provides a simple yet discerning starting point when evaluating storage I/O performance and when undertaking performance comparisons.
To further underscore the Top-Down Approach, both the hIOmon WIOP and the hIOmon WIOPF support historical displays, including those for an individual device or application upon a daily or overall totals basis. Of course, they also enable you to drill-down to an individual periodic time interval using the historical displays.
|Simple, Quick, and Easy|
Determining and assessing storage I/O performance can be complicated (if not perplexing). A primary goal of hIOmon is to help simplify such efforts without sacrificing accuracy, precision, and relevance.
Towards this goal, hIOmon has introduced features and capabilities such as those noted above (as examples, automatically-aggregated summary I/O metrics, the innovative hIOmon "Performance Threshold Range Metrics", and the DXTI metric).
The various configuration options also provided by the hIOmon "Disk I/O Ranger Display" and the hIOmon "I/O Trender Display" enable you to dynamically and easily control what gets displayed in accordance with your own particular needs and interests. Moreover, you can start with a minimal amount of details displayed and then include additional metrics to the display as needed.
In short, hIOmon enables you to determine and evaluate your particular storage I/O operation performance (especially the particular "I/O profiles" pertinent to your specific applications, workloads, and devices) in an efficient, timely, and cost-effective manner.
Bringing Transparency to Disk I/O Performance℠