Sebastian Stadil is founder and CEO of Scalr and founder of the Silicon Valley Cloud Computing Group.
Cloud computing has been on Gartner’s list of strategic technologies for the past five years and with good reason, the promise of self-service provisioning is accelerating application delivery and business innovation is strong. However, progressing toward this goal means making several critical decisions around each of the layers making up your cloud architecture.
With cloud being a vast and rapidly evolving ecosystem, the most crucial mistake an enterprise can make is to assume that all solutions that fall in a given category are equivalent. And while that realization can quickly make the architecture process overwhelming, we will break it down for you here with a set of questions for each cloud layer.
The three layers of technology that define the cloud stack – resource, platform, and management – each have their own options and key considerations. However, the ultimate business needs and goals should drive the decision process.
Enterprise resource layer
It’s highly likely that this foundation to the cloud stack that includes components such as hardware, storage, virtualization and network infrastructure, is already in place. And while the enterprise is likely already familiar with choosing the best-fit resource layer components for their organization, when constructing cloud architecture, this is an initial decision point. Enterprises must decide if they will work with a public cloud provider that manages this layer for them, if they will reuse their existing resources and manage their own resource layer in a private cloud, or combine the two approaches.
As such, the two critical business requirement questions enterprises need to ask themselves are:
- Are our planned cloud workloads sensitive and need to run on-premises? If so, private cloud will ensure these workloads meet regulatory, performance and/or security requirements.
- Are the workloads elastic with resource requirements varying greatly over time? Public clouds are best-suited for elastic workloads. Most often associated with development, test, and demo environment needs, public clouds will likely be the best, most cost effective choice for this type of requirement.
Cloud platform layer
This layer of technology automates resource provisioning by presenting an API that other pieces of technology can leverage and by translating requests made to that API into lower level commands that are sent to the resource layer. Just like the resource layer, cloud orchestration layers can be public or private.
Businesses assessing cloud orchestration platforms should ask themselves three important questions:
- Will cloud become a core competency? If so, the enterprise should consider OpenStack with its feature breadth and flexibility. Planning to make cloud a core competency does require that substantial resources are committed to achieve the goal.
- Does the organization value stability over development velocity? If so, CloudStack or a stable, packaged distribution of OpenStack such as Nebula One may be a better option, minimizing resources committed, with the trade-off being reduced feature sets.
- Has the enterprise already made a commitment to working with AWS and therefore API compatibility with AWS is critical? If so, Eucalyptus (recently acquired by HP) may be the best choice as it guarantees full AWS API compatibility.
Cloud management platform
The final layer before the application stack, cloud management platforms are the interaction layer between developers, IT, the business and the enterprise’s cloud infrastructure.
To determine the best cloud management platform, enterprises should assess:
- Is the enterprise seeking to manage a hybrid or multiple cloud environment? If the answer to this question is no, and an enterprise is managing one cloud, the cloud console that ships with their cloud orchestration platform, such as AWS Console or OpenStack Horizon may be a preferred approach. However, for enterprises seeking a single pane of glass to provision from and manage multiple clouds, a full-featured cloud management platform is recommended.
- Are business agility and cloud governance priorities? While cloud management is a crowded and ill-defined space, the range of available solutions is vast and enterprises should do their homework to ensure that their corporate priorities are reflected in the CMP capabilities. It is critical to not fall for the common misconception that all cloud management platforms feature the same breadth. Before issuing an RFP or selecting a vendor, it is highly recommended that an enterprise scope its governance and other operational needs.
- Is the goal of the move to cloud computing to accelerate development and innovation? If so, it is recommended that enterprises seek a cloud native cloud management platform, which supports a forward-thinking DevOps-centric approach to cloud management, as opposed to legacy cloud management platforms that simply layer traditional IT infrastructure provisioning workflow onto the cloud.
Enterprises should be careful in the architecture choices they make to not lock themselves into a solution that inhibits agility and their ability to dynamically manage cloud resources. Yet, defining a cloud infrastructure stack for an enterprise is a series of critical decisions that must complement each other. With each layer of the cloud stack, there are numerous options, and no two enterprises will share the exact same cloud stack as each needs to define its requirements and the solutions that best match those requirements and build the cloud stack that will translate into business success.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.