The evolution of technology can have significant influence on the future success, or failure, of companies that operate within an industry. Traditional communications service providers face this challenge today. The technologies driving this transformation are Cloud, Network Function Virtualization (NFV) and Software Defined Networking (SDN). Communications service providers (CSPs) look at the NFV model and see new ways to accelerate innovation, create competitive advantages, reduce cost and drive new business models.
Successful transformation requires that CSPs also evolve their operating model: evolving related roles, organizational structure, skill sets, processes and culture to reflect the reorientation of the company. We strongly believe that the “As-a-Service” cloud-based operating model is the right operating model to achieve the full benefits of agility, operational efficiency, faster time-to-market and cost reduction, as successfully demonstrated by early adopters like Google, Facebook and Amazon. These early movers are at a tremendous advantage. CSPs may be at risk if they don’t aggressively embrace As-a-Service capabilities.
Read this white paper for a thorough examination of the impact Network Function Virtualization will have on organizational structure and guidance on operating model transformation to NFV Infrastructure as-a-Service.
Enrico Montariolo is an Operations Architect with the VMware Global Professional Services and is based in Milan, Italy.
For seven years as an industry analyst, I studied top-performing IT organizations to figure out what made them best-in-class. And after studying 30 private cloud early adopters in 2011, I co-authored a book about how to deploy private cloud.
But after joining VMware last year, I’ve had the opportunity to spend six months working closely with VMware’s IT team to get an in-depth understanding of our internal private vCloud Suite deployment.
In this multi-part blog series, I’ll write about what I’ve learned.
Lesson learned – The most important thing I learned, and what really reframed much of my thinking about IT spending, is that VMware IT invested in our private cloud strategy to increase business agility. And that effort drastically lowered our costs.
Breaking it down:
1. We made a strategic decision to try something different.
Over the years, I’ve studied companies that use every form of squeezing IT budgets there is. But what happens with a “cut till it hurts” or a “cut until something important breaks” approach is that the primary objective of lowering IT budgets is often achieved. But it also leaves IT hamstrung and unable to meet the needs of the business. An unbalanced focus on cost cutting reduces IT’s ability to deliver. That in turn lowers business perception of IT value, which further focuses efforts on cost cutting. Define “death spiral.”
VMware didn’t follow that path when we decided to invest in private cloud. We justified our “Project OneCloud” based on belief that that the traditional way of growing IT capabilities wouldn’t scale to meet our growth objectives. We have doubled revenue and headcount many times over the last 10 years. The IT executive team had the insight to realize that a linear approach of increasing capacity by buying more boxes and adding more headcount would not support business needs as we double in size yet again. We are no longer a startup. We have grown up as a company. We had to try a different approach.
What really enabled our private cloud success was broad adoption. There is a commitment and investment in private cloud that requires broad adoption to justify the cost and effort. The promise of delivering IT services the same old way at lower cost didn’t drive adoption. What drove adoption was a new operating model focused on delivering and consuming IT as a service. Specifically, abstracting infrastructure delivered as basic compute, network, and storage as a service. Then designing IT services for specific groups of consumers that allowed them to get what they need, when they needed it. That included application stacks, dev/test environments, and any other business function that depends on IT infrastructure (almost all do in the mobile-cloud era). We strove to eliminate the need to call IT, and also eliminated tickets between functional groups within IT.
Ten different business functions — from sales, marketing, and product delivery, to support and training — have moved their workloads to the cloud. Many have their own service catalog with a focused set of services as front end on the private cloud. Many have their own operations team who monitor and support automation and process that are built on top of infrastructure services.
Carefully designing IT services, then giving people access to get what they need when they need it without having to call IT — is key to success.
3. Broad adoption drove down costs via scale economies.
We started with one business group deploying sales demos and put their work in a service catalog front end on the private cloud. Then we expanded onboarding other functional groups to the cloud. One trick – and that is to develop a relationship with procurement. Any time someone orders hardware within the company, get in front of the order and see if they will deploy on private cloud instead.
Make IT customers’ jobs easier. Accelerate their time to desired results. Build trust by setting realistic expectations, then delivering per expectation.
Three primary milestones:
Once we onboarded a few key tenants and got to ~10,000 VMs in our cloud, we lowered cost per general purpose VM by roughly 50 percent. With a new infrastructure as a service model that allowed consumers to “outsource infrastructure” to our central cloud team — and at a much lower cost per VM — word got out, and multiple other business groups wanted to move to the cloud.
Once we onboarded another handful of tenants and got to ~50,000 VMs in our private cloud, we lowered cost per general purpose VM by another 50 percent. We were surprised by how fast demand grew and how fast we scaled from 10,000 to 50,000 VMs.
We are “all in” and now on track to meet our goal of having around 95 percent of all our corporate workloads in private or hybrid cloud (vCloud Hybrid Service) – for a total of around 80,000 to 90,000 VMs. We expect cost per VM to drop another 50 percent.
So we set out to increase agility and better meet the needs of the business, delivered services that made IT consumers’ jobs easier, and as a result we dropped our cost per VM by ~85 percent.
Our private cloud goal was to reshape IT to better meet revenue growth objectives.
We transformed IT to deliver IT services in a way that abstracted the infrastructure layer and allowed various business team to “outsource infrastructure.”
Ten different internal business groups have moved workloads to private cloud.
Less focus on infrastructure and easy access to personalized services made it easier for IT service consumers to do their jobs and focus more on their customers.
A new operating model for IT and effective service design drove adoption.
Broad adoption drove down costs. By ~85 percent.
Below are links to two short videos of VMware IT executives sharing their lessons learned related to cost and agility. In my next post, I’ll talk about what I learned about a new operating model for IT.
—- Follow @VMwareCloudOps and @kurtmilne on Twitter for future updates, and join the conversation by using the #CloudOps and #SDDC hashtags on Twitter.