Virtana

Fig2 - Comprehensive Analytics Dashboard

Virtana is a SaaS based platform that combines high-fidelity baseline information provided by discovery, mapping, dependency analysis and profiling with analytics, to measure historical workload patterns and calculate the exact resource requirements in the various dimensions of CPU, memory, I/O, network and disk usage in public clouds. It then automatically formulates specific recommendations for each instance. This tool also allows users to apply additional customized constraints to control the outcome of the recommendations.

Virtana also leverages self-learning anomaly detection and pre-configured alerting policies to identify capacity bottlenecks, notifying administrators in real-time to avoid the risk of end-user impact. Unexpected capacity bottlenecks may happen because of a sudden change in application work, a fail-over, or an intended or unintended change in infrastructure configuration.

Virtana includes a capability to detect sudden increases in daily spend to avoid an end-of-month surprise. It automates the process of comparing cloud spending over time by correlating real-time data collected via APIs with detailed billing logs. The cost data can be grouped by meta-data such as attributes and tags, including related charges like data transfer and storage usage – all at an instance-level. Finally, it can account for the amortization of upfront spending commitments made to cloud providers that must be allocated to an individual instance based on usage. The result is a comprehensive yet simple report that can be set up once and regularly emailed to all stakeholders. This capability works on AWS, Microsoft Azure, and Google Cloud Platform.

More than 50 integrations (integrations in this context are systems that Virtana collects data from) are available, out of the box, for many of the leading cloud data, development, integration, and management tools. These integrations are offered with collectors, agents, and automation software. Additionally, a RESTful API allows you to read and write data programmatically. This means that you can perform a wide variety of tasks, including adding new integrations.

The first new module in 2021, will build on the existing Cloud Migration Readiness service. This uses detailed production workload information in private cloud and legacy on-premises applications from the WorkloadWisdom solution to analyze production workloads, create workload models, and then administer the generation of realistic simulated cloud workloads against any file, block, or object storage target via Workload Generators. It has a web-based GUI that is built for all user levels to provide an advanced networked storage performance validation solution that incorporates a sophisticated workload modelling methodology. New integrations with VirtualWisdom enable the capture and use of further high-fidelity information to help automate the process of discovery and profiling. While elements of consulting services are still required in this first release, there is a clear roadmap to providing a fully automated “know before you go” solution for customers wanting assurance about planned cloud migrations.

Further releases in the short to medium term are planned to provide a powerful Cloud Brokerage capability, and manage end-to-end visibility and management into a unified platform through a single pane of glass. We will revisit these areas when detailed functional capabilities are announced.

Business leaders have known for some time now that, in an increasingly digital-first world, the performance of customer facing applications can make the difference between business success and failure. Yet, as underlying infrastructures have become more complex, residing in both on-premises and cloud data centers, with diverse compute, network, and storage components from a wide range of vendors, understanding the impact and cross-correlation of events has become more difficult.

The growth in the use of public cloud, from Amazon, Microsoft and Google in particular, has enabled organizations to deliver and scale new applications and services faster than ever before. It also promised significant cost savings. This has not always been the case. Therefore, getting better visibility into, and control of cloud usage and cloud costs is becoming more important.

Today, few organizations maximize performance and availability of IT without any regard to cost. Likewise, few would ignore performance and availability completely in an effort to reduce costs. It is all about understanding and managing risk. Given the complexity of hybrid IT infrastructures, the volume and velocity of transactions and the speed of change, it means that balancing performance, cost, and risk is something that needs a deep understanding of your on-premises workloads and the interdependence of applications and their supporting infrastructure as well as automated systems with sophisticated artificial intelligence (AI), machine learning (ML), and strong analytics.

The Bottom Line

The ability to decode and understand where the best place is to run your business applications from a capacity, performance, and cost perspective, is now a critical business requirement. While Virtana stresses the importance of using data to drive more intelligent decisions, which we agree with, we believe that their legacy of skills and expertise in capturing and acting upon real-time, observed data from a wide range of sources remains a key differentiator. The latest developments Virtana are delivering, combined with that legacy, will place them in the forefront of the emerging multi-cloud migration, performance, and availability management market. To be competitive you need to embed intelligence in your infrastructure. The Virtana platform allows you to do that.

Figure 1 – VirtualWisdom Architecture

At the heart of the system lies the VirtualWisdom Platform Appliance. This is a purpose built, rack mounted appliance that persists and correlates real-time data from VMs, servers, SAN and NAS networks—with I/O metrics from Virtual Server Probes, Network Switch Probes, and “off-the-wire” from the hardware Performance Probes – forming a complete view of your end-to-end system between virtual machines, whether in the cloud or on-premises, and the storage environment as shown in Figure 1.

Where most IPM solutions sample data, VirtualWisdom captures information from software and hardware probes – at the microsecond level. By default, customers receive 16 days of per second data. Beyond that you are talking about capacity management rather than infrastructure performance. The huge amount of data ingested by VirtualWisdom 6.0 is fed into real-time dashboards that are easily configurable to provide drill-downs, relevant and understandable for both business executives and technical operations staff. Heuristics and machine-learning analytics provide cross-silo correlation of events to continually update and tune the effectiveness of the performance data and, additionally, provide the ability to rebalance VM workloads.

Existing customers well versed with the capabilities of previous versions will see a new iSCSI performance probe that complements existing Fibre Channel SAN and NAS performance probes. Meanwhile, agentless monitoring of storage arrays and hyperconverged systems have been enhanced to take in NetApp, IBM SAN Volume Controller (SVC), Dell EMC VMAX/PowerMax, Dell EMC VxFlex OS hyperconverged system (previously ScaleIO), Nutanix, and VMware vSAN.

Figure 2 – Example of application dashboard

In 2017 Bloor published a Spotlight; “Instrumenting the Virtualised Data Centre for Performance SLAs”. At that point there still seemed to be a lack of tools that could provide effective performance monitoring of virtualised and cloud-based servers as well as on-premises equipment. This release includes 3 significant new capabilities for monitoring those virtualised and cloud-based environments. Host operating system monitoring, and integration to monitor VMware vCenter, Microsoft Hyper-V and IBM Power-VM are covered by WisdomPack for enterprise compute, while WisdomPack for cloud compute offers agentless software integration for Host operating system monitoring within Microsoft Azure and Amazon Web Services (AWS).

A software-only deployment of VirtualWisdom 6.0 can generally be handled by customers themselves. However, a range of installation and support packages are available, and customers may want to avail themselves of one of these where hardware probes and the VirtualWisdom platform appliance are to be installed.

In an increasingly digital-first world the performance of customer facing applications can make the difference between business success and failure. As the quote from Gregory Phillips, who is a user of VirtualWisdom, shows, downtime and poor response times on applications can have a direct impact on the bottom line. Yet, as underlying infrastructures become more complex, residing in both on-premises and cloud data centres, with diverse compute, network and storage components from a wide range of vendors, understanding the impact and cross-correlation of events has become more difficult.

Ten years ago, Google discovered that an extra 0.5 seconds in search page generation time dropped traffic by 20%. They have spent a huge amount of time and money to ensure they understand exactly how their infrastructure performs….and their business flourishes. All businesses need that same sort of capability.

The Bottom Line

The ability to decode and understand the way your hybrid infrastructure interacts with your applications is now a critical business requirement. We are impressed with the strides Virtual Instruments have made in the last couple of years. To be competitive you need to embed intelligence in your infrastructure. VirtualWisdom 6.0 allows you to do that.