The IT industry is constantly evolving with visionaries leading the way for technology vendors to follow. Some innovations like storage tiering (e.g. EMC FAST) improve basic functions, making storage administrators more productive. Other innovations, like unifying management of converged infrastructure that combines compute, networking, and storage into a single, optimized computing solution, disrupt normal IT processes and administrators’ roles.
The data center is on the road from virtualization to cloud and beyond (e.g. the software-defined data center). Adopting new approaches like converged infrastructure with unified management can bring cloud services within reach. Unified management moves IT away from the time-consuming, repetitive manual provisioning processes used to provide infrastructure needed by new applications. Furthermore, data centers with converged infrastructure that replaces remedial actions with automation are better positioned to implement cloud models.
Thriving with Change
Disruptions are often perceived as negative because they mean change, but there are disruptions that can have very positive outcomes. Take disruptive innovation for example. Wikipedia states “A disruptive innovation is an innovation that helps create a new market and value network, and eventually goes on to disrupt an existing market and value network (over a few years or decades), displacing an earlier technology.” Examples of disruptive innovations abound such as personal computers replacing computer timesharing.
“What about disruptive innovation that just changes an existing market?”
Take Apple. They didn’t invent the smart phone. They changed how we use and interact with our smart phones, which completely redefined the market.
Potentially as impactful, but more disruptive, are the changes and new thinking needed to fully realize the promise of cloud architectures, and to successfully move to cloud services models (e.g. IT-as-a-Service). Many organizations initially embraced virtualization to reduce costs, but quickly learned that the concept of converged infrastructure is truly the foundation for them deliver cloud-based services. Other organizations virtualized their compute infrastructure but got stuck.
Facing the Facts
What is keeping these organizations from being more successful?
Many IT executives whom embraced converged infrastructure have IT departments that look at it as just more of the same. What the IT departments do not realize is that converged infrastructure provides an opportunity to streamline out-dated processes via a virtualized or abstracted environment that can be managed as one single entity. Human nature tends to resist change. Administrators want to continue to manage compute, networks, and storage as separate domains or silos. This approach perpetuates several challenges that keep IT organizations from ever getting to a services model and include:
- Many manual processes: Too many highly technical people are involved in provisioning infrastructure including the coordination it requires, between silos. This approach wastes time, money, and talent because it relies on many repetitive tasks and is not the best use of high-salaried personnel.
- Multiple provisioning steps: IT employs manual provisioning via scripts, with multiple hand-offs between departments, to provision compute, network, and storage, and to integrate with virtual stacks like VMware vCenter. This approach is prone to human error and fixing mistakes becomes costly because it takes time to determine issues, remediate, and re-start processes.
- Misuse of resources: There is also a tendency to over-provision resources because actual needs are not known in the early stages of deploying a new application. This common practice results in infrastructure components not getting released at a later time, and valuable resources remaining unavailable for other applications.
Example: Multiple Provisioning Steps in a VMware Environment
(subset—there’s much more; do once and repeat, repeat…)
As a result of these multiple provisioning steps it takes many weeks to deliver infrastructure for new applications, which can delay the rollout of new services and potentially negatively impact the bottom line. Making matters worse yet for IT is when the line-of-business (LoB) goes to a public cloud for infrastructure services to run their new applications, because internal IT processes take too long.
What can be done—to change?
IT managers need to make the hard decisions, such as changing how setting up infrastructure for new applications gets done. Making the changes now will position a data center well for moving to cloud architecture.
These changes include deploying new technologies to:
- Create services catalog: This step encompasses the creation of ready-to-provision infrastructure services (compute, network, storage), which can be placed into a catalog for selection when an application is ready to be deployed. IT creates templates that define the maximum resource constraints and default resources that can be used to provision services. This setup requires close coordination between LoB and IT for best results.
- Automate processes: Provisioning needs to change from silo-based operations, utilizing manual scripts, to a completely automated provisioning process up and down the stack. Once done, IT can dynamically respond to new or changing business needs. IT can shift unused resources between services, and add or remove infrastructure components to an already provisioned service with just a few clicks. Converged infrastructure management solutions are out there like EMC Unified Infrastructure Manager which is the most mature in its third iteration.
- Better leverage experts: With a services catalog and automated provisioning in place, technical experts within the silos still determine the resources and services needed, but pass on the provisioning tasks (now completely automated to eliminate the manual provisioning work described above) and infrastructure services management to a general administrator.
Get Great Results
Moving to converged infrastructure with centralized provisioning management capabilities enables IT to create and provide a catalog of predefined “infrastructures”. For those data centers moving to more sophisticated cloud services models, IT can proceed to empower their LoB users with the ability to request these “infrastructures” through self-service portals. Organizations can dramatically improve their responsiveness and ability to delivery cloud-based services with converged infrastructure, including centralized provisioning. Consider Nordic IT infrastructure services provider ATEA, for example. ATEA went from a 2-week infrastructure service delivery time down to 30 minutes, and now has an SLA to deliver a virtual machine within 5 minutes. However, by using EMC Unified Infrastructure Manager for their converged infrastructure (VCE Vblock), they can deliver a virtual machine in one minute.
Example: Access Infrastructure Services Catalog
(complete process – provision once, use many)