GUARANTEEING the stability and performance of mission-critical applications for optimised IT performance is a vital but impossible task without the support of analytics providing accurate insight in what’s really going on.
With the down sides of escalating costs due to overprovisioning, little clarity or assurance on performance or the ability to adequately meet SLAs, a reliable prediction capability which can offer insight into the health and performance of the data centre is undoubtedly the key to not only cracking these issues, but a means of future proofing systems to ensure the business can deliver on its promise to customers.
When we look at the latest developments in terms of analytics applied to obtaining performance intelligence from the entire IT infrastructure, we enter the category of Infrastructure Performance Management (IPM). This is a relatively new capability developed for datacentres that presents end-to-end view of the entire IT infrastructure by enabling the continuous capture, correlation and analysis of system-wide performance, utilisation and health, in real-time.
Developed directly from the expertise gained from working with hundreds of enterprise organisations, the granular level of analytics now available from IPM is designed to provide data and insights aligned exactly to business need and even various user roles. Applied analytics turns data into real answers to help IT quickly address common performance management requirements, such as workload balancing, event investigation and trend correlation across the environment.
How infrastructure performance metrics benefit the organisation
Bespoke datacentre requirements
IPM metrics can deliver meaningful
recommendations in the context of an
IT environment’s own unique demands
and workflow to help guarantee SLAs and
optimum infrastructure performance.
Performance intelligence
The latest IPM technology provides
insight at a glance, allowing IT to better
understand the complex inter-connectivity
and relationships between applications
(legacy or new) to enable a high-level
knowledge of resource utilisation, health
and performance, and to put an end to
expensive overprovisioning.
Predicting performance
Metrics in action: the predictive quality
of Infrastructure Performance Management
indicates a shift from reactive
troubleshooting to proactive management
as users can view history and trending
information so they can take action before
an issue becomes a production-impacting
event.
Real-time reports and wider usability
An IPM solution can produce live
reports providing intuitive and dynamic
visualisation of millions of wire, machine,
and analytic metrics across the
performance, utilisation, and health of
devices based on workflow. For the
first time, information can be presented
to each user or group (be they CIO, IT
manager, server or storage team) in a
format relevant to their job function, such
as user-friendly bar charts, trend graphs,
time based comparisons and histograms
at an unmatched level of granularity, all
of which can be used to support strategic
IT decision making and collaboration.
Advanced applied analytics tools can execute tasks in just a few seconds, which prior to IPM analytics, would have taken hours or even days to achieve. The forensic level of detail obtained through IPM metrics enables organisations to identify the exact location of any potential bottlenecks and fix them so these never need reach customers.
“Good enough today” won’t work tomorrow
According to a Forrester study conducted in September 2014*, more than 55% of the decision-makers surveyed identified advanced analytics as having a high impact on lower frequency and shorter length of service outages, customer satisfaction, as well as higher IT staff productivity and job satisfaction.
The report states that organisations should prepare for a ‘workload-centric’ infrastructure to ‘achieve a dynamic environment of better hardware utilisation, increased workload mobility, and faster application deployment to keep up with the velocity of today’s increasingly agile development and operational environments.’
“With advanced analytics, enterprises can quickly see how well the infrastructure is transporting applications and how well those applications are performing in the context of the end user experience, whether internal or external to the business.” Forrester
According to the study, changes, patches, upgrades, component failures and software limitations/misconfigurations were identified as the principal issues contributing to performance problems for infrastructure and operations decision-makers. Additionally, the lack of end-to-end visibility, predictive analytics, and correlation between data from different infrastructure components were cited as the main performance data problems in their organisation. The research also indicates that many of these enterprises even slowing down progress to avoid performance issues.
As the health of performance in IT infrastructures depends on the complex interaction of thousands of moving parts, each generating various performance metrics, problems have grown beyond the capabilities of the enterprises and their best engineers.
VirtualWisdom4: Applied Analytics
VirtualWisdom4’s best-in-class capabilities for deriving definitive and actionable insights from these comprehensive metrics deliver tremendous value across the Infrastructure Performance Management domains:
Proactive Real-time monitoring
VirtualWisdom4 provides early
identification of emergent health,
utilisation, and performance issues,
enabling rapid implementation of
corrective actions to improve performance
and availability and reduce operational risk.
Deep diagnostics
Detailed, time-series correlated data helps
to identify the root causes of even the
most sporadic and complex performance
and availability issues, so they can be
definitively resolved and prevented from
reoccurring.
Infrastructure optimisation
Comprehensive insights into the
relationships between workload,
performance, and utilisation that supports
the optimal alignment of application
demands with infrastructure capabilities -
deliver the required performance at the
lowest cost and highest availability.
Performance-based service level
Agreement (SLA) Management
The ability to measure
and generate
detailed histogram reports on I/O
performance for every single exchange
facilitates the implementation of true
performance-based SLAs, enabling tighter
and more transparent alignment between
infrastructure, application, and line-of-
business groups.
Performance management metrics
The VirtualWisdom4-compatible
performance probes introduce a set of
new metrics that
help users quickly
identify and resolve performance
problems. For example “host or array-
side write delays” tells the user how
much of the system-wide latency was
attributed
to waiting on the host versus the write latency introduced by the
array.
Stabilisation of the system-wide infrastructure platform in order to provide customers with the very best resilience and performance guarantees is a top priority for organisations. Virtual Instruments’ Applied Analytics focuses on the critical findings that exist beyond the raw data to provide the built-in algorithmic logic necessary to derive that insight and present it right - all within the VirtualWisdom4 user interface.
Implementing an IPM programme using applied analytics will put CIOs, IT directors and decision makers ahead of the game, armed with the knowledge that expensive flash procurement (for example) will be a viable option to improve performance for
that organisation’s unique requirements, because the CIO will know from the intelligence reports whether or not
there are other problems affecting performance and that his purchase will
really make a tangible difference to performance that will be reflected in
ROI figures, rather than a costly mistake.
