After leaving office and serving as the Vice President of the United States, Al Gore earned the Nobel Peace Prize in 2008, for his efforts and dissemination of thought leadership surrounding man-made climate change, and laying the foundation for initiatives to counteract the impact.
Al Gore was able to play such an instrumental role, as he understood both business interest and public policy. With his leadership and experience, Al Gore focused efforts on what he referred to as ‘Grid Parity’, during a TED Talk. “Grid parity is understood as that line, that threshold, below which renewable electricity is cheaper than electricity from burning fossil fuels,” he explained.
Al Gore understood that once renewable energy was on that line or lower, then business interest would turn to the renewable energy source and this would be the game changer for the Environment.
Parity is the operative word, whilst many organisations are already taking advantage of cloud computing, there are still many that struggle to build their business case and find that ‘Parity’ to support cloud adoption. There are key factors that influence the business case and whilst these are common across each industry, the cost values are unique to the organization and its operating model.
Hybrid cloud is the most common deployment model for the enterprise and this model recognizes that not all workloads that exist, are suitable for cloud adoption and workloads should be assessed for cloud suitability. For example, a workload with a fixed computing load, very low network latency and tightly coupled integration, may be better suited for an on-premise environment. On the other hand, workloads with dynamic variability and easy access over the internet are better suited for cloud.
A business case or economics study will typically compare on-premise workloads (‘As is’) to cloud (‘To Be”). When performing a comparison, consideration is required for the additional value that cloud provides over on-premise workloads. This approach ensures an Apples-to-Apples comparison and avoids the pitfall of Apples-to-Cloud.
Identifying the on-premise costs is a more simplistic exercise as the CIO and organization already have visibility of both capital (CAPEX) and operating expenditure (OPEX). Determining the value that cloud provides requires more consideration on the impact to the specific organization and how it will better enable its Business objectives.
There are five characteristics of Cloud as defined by NIST and this is the best place to start when considering the cost-value differentiation between on-premise and a Cloud Service Provider:
- On Demand Self Service
- Broad Network Access
- Resource Pooling
- Rapid Elasticity or Expansion
- Measured Service
On Demand Self Service typically provides the Service Catalog, Automation and Orchestration components. The cost value in this characteristic stems from the ease and speed of provisioning, and thereby accessing services without human intervention. The ability to rapidly provision services for some organizations has more $ value than others.
Broad Network Access is the most straightforward of the five characteristics, where services are available over some form of Network (public or private). As a global organisation, selecting a Cloud Service Provider with a global data center footprint and high speed network will be more cost effective than provisioning it’s own infrastructure in every local market.
Resource Pooling and Virtualization have been around decades. IBM invented virtualization in 1967 to increase utilization of the mainframe. The IBM “Supervisor-Code” had a one to one relationship between the OS and hardware. IBM introduced a “Hyper-Visor” above the “Supervisor” as a means to abstract the hardware from the OS and increase utilization of the mainframe. In 2001, VMWare introduced “Elastic Sky X” or ESX to take advantage of the growing x86 processing power. The shift towards multi-tenancy of physical hardware components gave birth to cloud computing as we know it today. Many organizations have already implemented some form of resource pooling and virtualization to increase on-premise infrastructure utilization.
Rapid Elasticity is a fundamental cloud characteristic. Some of the most suitable workloads for cloud adoption are those that are variable in nature. With Cloud Service Providers offering services by the hour, taking full advantage of this offering requires dynamic workloads with changing demands. Paying only for what is consumed, lends itself to those peak periods when a retail Organization requires more power during the holiday peak seasons than off peak, or when an organization’s major processing is conducted during office hours with limited activity in the evening or vice versa. Serverless architectures such as OpenWhisk are set to transform the way applications are executed through event driven apps that spin up and down based on demand and triggers. In the Serverless Architecture, application actions run automatically and only when needed, thereby significantly reducing Infrastructure footprint until the action call is made.
Measured Service or Metering drives the desired behavior with greater transparency. Organizations often consume unnecessary resources for two primary reasons; 1. Belief that spare resources are available due to over provisioned infrastructure (headroom for capacity growth), 2. Lack of Showback or Chargeback mechanism (buffet syndrome). Consider your own consumption behavior at a buffet versus a la carte. In a buffet you’ve already paid a set amount, so you consume the lobster, shrimp, fillet steak and everything else the stomach desires. For a la carte, you consume more sparingly and based on the contents of your wallet, “Chicken Caesar Salad please”.
For an Apple-to-Apple comparison and a comprehensive economics study, consideration must be applied to these five characteristics to address the qualitative benefits (Flexibility, Speed, Agility, Innovation) albeit, using intelligent assumptions to quantify them.
A number of cost dimensions should be explored for an Economics Study:
- Components (e.g. Facilities, Hardware, Software/Workload, Connectivity).
- Labor (e.g. FTEs, Upgrades, Expansion, Migration, Burdened labor).
- Quality of Services (e.g. SLAs, Agility, Speed, Flexibility, Innovation)
- Net Present Value (e.g. Cost of Borrowing)
The first two cost dimension are typically covered by most economic studies/business cases, with the last two that are most commonly overlooked.
The Quality of Services is typically overlooked due to the qualitative nature. However this dimension can determine if an organization remains competitive or even continues to exist. Deeper analysis on the value derived from improved speed of innovation can ultimately achieve ‘Parity’ and beyond for a more compelling business case.
Moving from a CAPEX to OPEX model provides a financial benefit to the organization. An organization has to raise capital for new infrastructure and applications. This capital is an upfront investment with a ‘cost of borrowing’ associated to it and a ‘Net Present Value’. The organization could make better use of investment funds to launch new products and services.
All these factors should be considered and understood when assessing the economic viability for cloud workload adoption. The Apples-to-Apples comparison is a more reflective assessment of real value and convenience that cloud based services will deliver to each unique organization. A one size doesn’t fit all, and this is the main reason for the gap between same sector organizations that have already adopted cloud and those that are still preparing for the journey.