Cloud 101CircleEventsBlog
Register for CSA’s free Virtual Cloud Trust Summit to tackle enterprise challenges in cloud assurance.

The Evolution of Private Cloud Computing and Shared Responsibility

The Evolution of Private Cloud Computing and Shared Responsibility

Blog Article Published: 09/15/2021

Written by Vishwas Manral, Founder at Precize Inc & Fellow at Cloud Security Alliance.


Summary:

Cloud computing has changed over the last 10 years – the Private Cloud has undergone a big change too. This blog captures the evolution of the Private Cloud with a focus on the shared responsibility model. A previous blog talks about the different service models as they have evolved in the public cloud.

Background:

The National Institute of Standards and Technology (NIST) provided a definition of cloud computing in 2011 comprising four deployment models – Public Cloud, Private Cloud, Hybrid Cloud and Community Cloud.

The Private Cloud deployment model is defined as cloud provisioned for a single organization and owned, managed and operated by the organization or a third party.

In most cases the model defines a single party managing the complete deployment.

With public cloud vendors aggressively expanding into the private cloud space, the responsibilities for organizations are changing. There is now a clear movement of Public Cloud stack into the Private Cloud.

This blog is intended to provide an overview and guidelines for shared responsibility for the evolved Private Cloud as it stands in 2021, especially in comparison to responsibility for the traditional Private Cloud as defined by NIST.

Key Change Driver:

Bringing new value to the marketplace is now a competitive differentiator and market leaders are now organized to deliver on innovation quicker again and again.

The centralized nature of the cloud enables quicker development and delivers cost efficiencies. However certain applications cannot be moved to the Public Cloud and data/applications have to be kept on the Private Cloud due to compliance, latency or for data jurisdiction reasons.

Private Cloud is evolving to a model where the Public Cloud stack is now being extended to the Private Cloud and cloud vendors are now taking additional responsibility in the Private Cloud. This blog talks about some of the newer service models for Private Cloud and the distribution of responsibility between the application owner and platform provider.

Resources in the Private Cloud are managed just like resources in the Public Cloud and give a unified management for both the Private and the Public Cloud.


Managed Hardware Infrastructure as a Service (HIaaS):

In this model, the infrastructure, software service stack, APIs and service tools are provided by a Public Cloud vendor to be installed on premise of a customer. As the infrastructure and the service stack is managed by the cloud provider, the organization has responsibility only for the application stack running on the infrastructure. An example of this service is AWS Outpost or Azure Stack Hub/HCI/Edge where the infrastructure is shipped, installed, and managed by Public Cloud vendors.

The Private Cloud in this case can be the edge site or a data center. Based on the service stack provided the service provider could take additional responsibility, such as taking responsibility of the service layer.

Managed Control Plane as a Service (CPaaS):

Managed Kubernetes is the most widely used Service layer. Managed Service Control Plane is a service (CPaaS) provided by most cloud providers for Public Cloud and the same stack is extended to the Private Cloud. In this case, the Service (k8s) control plane is managed by the service provider with some configuration optionally provided by the application owner. However, the infrastructure is managed by the organization and not the public cloud vendor. Bringing up the data plane and managing it is done by the application owner, with some help from the service provider. Examples are AWS EKS Anywhere, AWS ECS Anywhere, Azure Arc and Google Anthos.

As the cloud vendor is managing the control plane in this case, one could use an IaaS service from AWS and run Google Anthos on it.

Conclusion:

In summary, the future landscape of applications is highly hybrid or multi-cloud. Enterprise cloud computing will include a vast variety of infrastructure, platforms and APIs including serverless and server apps, on-premise and cloud.

With the renewed focus on Private Cloud, the number of service models for Private Cloud will greatly increase in the coming years all towards the benefit of customers.

I invite you to connect with me on Twitter @vmanral or Linkedin to share your thoughts and provide your input.

Share this content on your favorite social network today!