Case Study / ROI Training

“We used to provide our own hardware for all hands-on classes. With our global customer base expanding and looking for training on more cutting-edge technologies, we need scalability - where provisioning and shipping our own hardware does not always make sense. This is where Ravello has really helped us.”
Dave Carey
President & Co-founder, ROI Training

ROI Training - a leader in enterprise training

ROI Training is a leading provider of technical, financial and management training for enterprises around the globe. With a modular approach to course content, ROI customizes the training to suit to the needs of the organization. Featuring over 150 course titles, their technical training spans over a wide domain - from Google Cloud Platform to OpenStack to Big Data - all delivered by experts. Since its inception, ROI has served a large and loyal customer base - comprising of hundreds of customers ranging from midsize to Fortune 100 companies.

ROI Training's need for training lab environments

Over the years, ROI Training has found that the most effective way for the students to learn is through practice. Towards that end, all technical courses offered by ROI Training have lab components where the students get to work on systems that mirror real world deployments. Training 1000 students per month, ROI Training needs a lot of lab environments.

Lab environments for some of these courses (e.g OpenStack) are very sophisticated - requiring 11 virtual machines and L2 networking for a variety of functions - compute nodes, neutron, controller, swift, router etc. During the course, students start with 4 nodes, and progress their way to configuring 11 nodes connected through 4 physical networks and tied through routers. This ‘hands-on' approach helps them clearly understand how the different technologies of OpenStack work together. To build and run an OpenStack environment as a part of the training requires 40 vCPUs, 120 GB RAM, 350 GB of disk space for each student or 800 vCPUs, 2.4 TB RAM, 7 TB of disk space for an average class.

ROI Training Management Network

ROI's requirement - provide effective learning to its trainees

To provide an effective learning experience, ROI wanted a production scale OpenStack lab environment for each of its trainees. Its requirements for training lab environments included:

  • No CapEx investments - ROI Training experiences variable workloads depending on the number of classrooms in-flight. They wanted to avoid investing in hardware to run training labs, and cost-effectively leverage the utility of the cloud.
  • Scale on-demand - Each student in an OpenStack course needs a large amount of compute resources (40 VCPUs & 120 GB RAM). With an ever-growing number of classes, ROI needed a platform that could scale to accommodate their needs without impacting the training lab experience.
  • Lab templatization - To avoid a recurring configuration effort every time they offered a course, ROI Training needed a way to templatize their lab environment, and use it to create student training environments.
  • Ease of setting up virtual networks - To provide an easy way for their students to put key concepts in action, ROI Training needed the capability to easily create virtual networks.
  • Neutron networking support - OpenStack comes with its own networking called ‘Neutron'. ROI needed support for promiscuous mode to be able to setup the Neutron networking.
  • Console access - Having console access to each of the machines in the lab setup was paramount to be able to configure the systems.
  • Geographical reach - ROI has customers in every continent. They needed a solution that allowed them to create lab environments at locations around the globe - to be able to provide a ‘local' end-user experience.
  • Usage-based costs - To reduce the overall cost of training labs, ROI was looking for a strictly usage-based cost structure.

Before moving to Ravello, ROI explored several options. They looked into using customer's resources to setup the training labs. However, given the time it takes to set up the lab environments and third-party dependency for hardware, they quickly moved away from this model. Next, they looked at providing their own hardware & software for lab environments at each of the training locations to save on setup time. Shipping hardware each time can be quite expensive. Also, it was not scalable to ship hardware that could accommodate 9000+ VMs that ROI Training needed during an average month.

Ravello - a perfect match for ROI's needs

ROI tried Ravello for their OpenStack labs and were extremely happy with the rich functionality that the platform offered, and the customer support they received. ROI now uses Ravello for a large number of their technical courses.

ROI Requirement How Ravello delivered?
No CapEx investments Ravello's solution allowed ROI Training to create their training labs on Google Cloud and AWS, eliminating the need to invest in their own hardware resources.
Scale on-demand With Ravello, ROI Training could create as many resource-intensive training labs as needed without having to worry about capacity restrictions. Since Ravello runs on AWS and Google Cloud (Tier 1 cloud providers), there is never any shortage of capacity, quota, and overage concerns.
Lab templatization Ravello's blueprint feature allowed ROI to easily create lab environment templates for each of their courses that could be used to spin-up copies of the environment for each of their students
Ease of setting up virtual networks Ravello's intuitive GUI and built in network services such as DNS, DHCP, IP Filtering, virtual switch and virtual router enabled users to easily create and modify sophisticated virtual networks.
Neutron networking support Ravello's software defined networking overlay exposes a clean Layer 2 and Layer 3 interface that supports promiscuous mode and allows OpenStack Neutron to run unmodified.
Console access With the functionality built-into Ravello's HVX hypervisor, ROI was able to get console access to each of the compute nodes - otherwise not possible on public cloud
Geographical reach Ravello supports most of the AWS & Google cloud regions, allowing ROI Training to spin up training labs around the world, closer to their customer locations.
Usage-based costs Ravello is Software as a Service (SaaS) offering, and ROI only gets charged based on actual service usage.

Using Ravello, ROI has been able to build some very sophisticated multi-VM blueprints for their training labs.

ROI Training lab on Ravello

Benefits realized with Ravello

ROI has realized many benefits through adoption of Ravello as the platform of choice for its lab environments.

Up-to-date curriculum - Due to frequent updates associated with cutting edge technologies such as OpenStack, the trainers have to rebuild the lab environments with every new release. With Ravello's modular approach to creating an application environment and ease of network setup, ROI has been able to keep its curriculum up-to-date with reduced effort.

Better learning experience - Using Ravello, ROI Training is able to provide students large scale labs that mimic the real world production deployments - enabling an effective learning experience. In many cases, students have used Ravello to recreate issues that they have seen with their production setup and ask the instructors for help.

Reduced time for lab setup - Being able to automate the repetitive lab setup tasks using Ravello's blueprint feature and training portal, ROI now spends less than 15 mins to setup the environments before the starting the course. Before moving to Ravello, the same set of tasks used to take 2 days.

Cost savings - ROI Training has seen a 65% reduction in their costs to set up training labs by using Ravello. With a usage based pricing, ROI only pays for the time their classroom lab environments are running. As a result, they have been able to run 15 concurrent labs for the same costs that it took them to run 2 environments using hardware.

Prompt issue resolution - ROI Training was amazed with the excellent customer support offered by Ravello's support engineers. Everytime ROI Training hit a ‘bump-in-the-road', Ravello's support engineers were quickly on a call to help troubleshoot the issues. In Arthur's own words, "Irrespective of whether it was a Ravello problem, OpenStack problem or a configuration error on our part, Ravello support engineers were willing to help. They were with us on calls night or day to ensure that the problems were fixed promptly. This willingness to help is something I have not seen anywhere else."