Ample evidence of the cloud adoption from small businesses to large global enterprises reflects the growing significance of its role in helping businesses stay relevant and agile. Cloud providers offer services with varying levels of abstractions that range from business process as a service (BPaaS) at the edge to infrastructure-as-a-service (IaaS) at the core, with multiple intermediate layers such as software-as-a-service (SaaS), and platform-as-a-service (PaaS).
Consumption to hold the key to cloud
Availability of a service on cloud must, at the least, be based on three principles or three Is namely - infinite, instant and independent. Infinite implies that for all practical purposes, there should not be any constraint on the quantum of the service being consumed, so long as the service is released back to cloud source as soon as it is no longer needed. Instant implies that the time required for provisioning of cloud-based service should be significantly low as compared to the duration for which the service is used. Lastly, independent implies that any consumption-related information that is required to order the service including pricing and features should not be hidden from a legitimate customer. This brings us to a few questions when we consider the “consumption” of these cloud-based services.
- Are the cloud providers able to provide best possible ‘pay-as-you-use’ options for organizations to consume their services?
- Do the cloud players offer flexible options at the cost of quality and reliability of services?
- Are businesses really satisfied with the flexibility of choices in terms of how they actually do or do not consume the services?
But before one attempts to look deeper into the nuances related to consumption, it is important to appreciate that the commoditization of computing has its beginnings in the core infrastructure (a.k.a. IaaS) offerings, made possible with the virtualization of computing. Hence a “pay-as-you-use” arrangement, signifying a “pure” consumption model, has traditionally been strong as far as IaaS offerings are concerned.
Another important aspect is the dominance of “public” cloud infrastructure as large cloud players started investing early into mega cloud infrastructures creating farms after farms of servers. This has left the cloud offerings divided into “public” cloud-led IaaS offerings with “pure” consumption model, and the rest of the cloud offerings with pseudo-consumption models.
While the public cloud-led IaaS has provided early thrust to cloud adoption, a sustained cloud adoption is dependent on maturity of consumption models of private clouds led by offerings like PaaS, SaaS and BPaaS, as businesses increasingly will be looking for complete offerings, and not just infrastructure on cloud. Also, owing to the growing concerns about data security, organizations are likely to opt for a private cloud arrangement to some extent. For instance a recent survey from IDC found businesses to be moving away from public cloud due to security concerns.
Consumption patterns grow as one moves from core to the edge
As noted earlier, IaaS offerings are more mature in terms of the consumption models that they offer. The services are very granular and can be dynamically configured or changed as per specific need. The consumers have the flexibility to pay for the services based on their actual consumption like the use of compute, storage and other infrastructure-related resources.
However, for higher level of cloud services such as SaaS or BPaaS, the consumption options available still seem to be quite restricted. For example, the HR cloud players typically offer “per employee” model, Finance and Procurement cloud players offer “per Invoice” or “per PO” model. While the cloud players do provide consumption options based on usage such as number of active users and key business transactions completed., such options may not truly match with unique consumption patterns of an organization. Organizations could potentially benefit if the right usage model is offered that matches with their consumption pattern.
There seems to be some light at the end of the tunnel
The emergence of “serverless” computing and related cloud models such as function-as-a-service (FaaS) has enabled organizations to run a function (code) without having to provision for any servers and pay only for the compute time which was actually consumed. FaaS, with its serverless computing architecture, has taken the PaaS model to a completely new level owing to the new consumption model approach that comes with zero idle footprint.
Cloud providers need to create “Intelligent Consumption Models” that use dynamic units for measurement based on the actual consumption patterns. AI can be used for prediction of specific patterns of consumption and to efficiently provision required amount of resources needed to support the services. For instance, HR SaaS consumption model can be based not just on the number of users but also other usage aspects like functionalities used, specific features used, frequency of use, number of processes completed and even the specific time periods when they are actually used. The consumers will then truly have the flexibility to consume services as they wish and pay as per the same.
Will the move from “pay-as-you-use” to “pay-as-you-consume” result in the next wave of cloud adoption?