Automated Workload Deployment, Governance, Auditability & Portability on Hybrid Clouds for Digital Transformation
Global businesses unquestionably need to improve their business agility through Digital Transformation, accelerated by the improved technology agility in terms of delivery, governance, auditability and portability of the business workload solutions, made possible by cloud services, the Internet of Things, big data and SaaS platforms.
In the broadest sense, this is called the Digital Transformation of the Enterprise – where more and more operational and contextual data is gathered, made available through broadly accessible workloads, and analyzed to unearth problems and opportunities on a regular basis. A digital enterprise has agile workloads, so it can respond quickly to what is learned and is able to use the best resources without the risk of being locked in to run these workloads on. These challenges are compounded by the growing complexity of the workloads that automate the business; by the growing need for workloads portability to be deployed to different clouds and workload availability to meet the growing global audience of employees, business partners, customers and the general public needs; and by the ever-growing importance of security, privacy and regulatory compliance.
Large scale, and dynamic workloads require a new approach to automation, which can’t be achieved with today’s tools that depend on risky manual processes and expert intervention. The next generation of automation must capture workload requirements and lifecycle semantics in a set of abstractions to provide the fundamental operations needed to manage applications throughout their lifecycles. This workload -centric perspective raises the workload to the highest level of importance and views infrastructure as subservient to the workload. In an workload -centric world, the value of the infrastructure is determined by how efficiently and flexibly its capabilities can be used to fulfill the requirements of the workload and provides the customers the flexibility to use the best infrastructures without the risk of lock-ins.
The most profound business automation is enabled by a collection of business workloads, many of which are intrinsically complex, reflecting the interdependencies, necessary policy-driven constraints and KPIs of the business processes. Deploying, governing, auditing and ensuring the security of these workloads is in turn a complex process, during which the workload is:
1. Sized (capacity constructed to match the anticipated loads)
2. ‘Built’ (constructed from component modules and libraries) along with the provisioning of the various operating system and middleware software needed to support the application operation
3. And finally ‘integrated ’ to the specific infrastructure system (server, storage and networking) hardware needed to run the application and integrate it with other IT assets.
Building workloads that perform well and are reliable is critically important, so workload deployment is a thoughtful and careful process, historically characterized by the collaboration of a set of consultants and experts, each of whom has intimate knowledge of only part of the process – workload details, storage details, network details, server details, operating system details, etc. The knowledge of these experts is about the dependencies, constraints and KPIs that define the objective specifically.
The building process starts with the ‘outermost workloads’ that define the functions to be used. Workload dependencies determine all the modules and libraries those elements need, which leads to further dependencies that the modules and libraries need, and so forth. Applied iteratively, this leads to all the hardware and software elements required to constitute the entire workload system. The constraints and KPIs are additional facts about the desired workload system that must be true (e.g., building on a specific family of hardware, or security rules) in addition to providing all the dependent resources that are needed. The use of formal blueprint models serves the dual purpose of formalizing the deployment process and also generating a formal definition that can be used as input by automated tooling or for programmable infrastructure.