Cloud computing is all the rage these days and there is no shortage of content on the definition of a cloud.  While the concept of a “Public Cloud” is fairly well defined and widely accepted, the concept of a “Private Cloud” seems to be up for debate.  I have been part of few conversations where the term “private cloud” is reduced to a meaning less mnemonic and an oxymoron.  There have been numerous arguments in the blogosphere about the validity of private clouds and some has gone to the extent of saying that no such thing called private cloud.

These arguments have merits depending on how one looks at the definition of private cloud.  I think most of these arguments against the private cloud really miss the mark on what really makes a cloud, well, a cloud.  Accordingly to NIST:

Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction

Keeping that definition in perspective, a private cloud is really just a cloud deployed inside an organization’s boundary (within the firewall) and is managed by internal employees. This refers to cloud computing where an organization virtualizes and utilizes its own IT infrastructure in a managed services model to deliver IT resources to internal application owners over the internal private network.   I would argue that a cloud is a cloud, irrespective of its deployment model and whether it is available for public use or private use.  The characteristics and benefits inherent to a cloud still apply to a private cloud as much as they do to a public cloud.

Private Vs. Public Clouds

Let’s look at some of the essential list of attributes that define a cloud and compare public and private clouds based on these attributes.

  • Elasticity: The ability to dynamically provision and de-provision the computing capacity as needed
  • Virtualization: The ability of applications to be de-coupled to from the underlying hardware
  • Metered Usage: The ability to automatically control and optimize resource use through the use of a metering capability appropriate to the type of service
  • On-demand Self Service: Ability for a consumer to unilaterally provision computing capabilities, such as server time and network storage, as needed automatically without requiring human interaction with each service’s provider.
  • Resource Pooling: Ability of the provider’s computing resources to be pooled to serve multiple consumers using a multi-tenant model, with different physical and virtual resources dynamically assigned and reassigned according to consumer demand.
  • Security: The ability to secure the data within the cloud and access to the cloud itself
  • Manageability: The ability to effectively manage (start/stop/migrate/expand/shrink/etc) the different server and application instances in the cloud

Private Vs. Public Clouds

As can be noted from the comparison above, there is not much difference between a private cloud and a public cloud from the perspective of the cloud attributes, other than a private cloud theoretically be more secure than a public cloud.  As a result, private clouds provide the same (sometimes more) benefits as a public cloud but for the economies of scale that is inherent to public clouds (more on this later).  In addition to inheriting all the benefits of a public cloud, private clouds also make the IT department of an organization more agile and responsive to business needs.  Business units no longer need to ask for more servers and wait six months to get them procured, installed and configured.  A private cloud, whether it is built in internal data center or built and hosted by a third-party, provides an organization with a dedicated pool of IT resources.   This pool will no longer be thought of in terms of number of servers, but rather in terms of capacity – the number of virtual servers, virtual workloads or applications it can support.  Basically, it allows IT departments to deliver infrastructure, platforms and applications in an easily managed, easily scaled and easily billed service architecture. In addition, this architecture will not be siloed to different departments or divisions, but instead can be managed as a holistic resource across the entire organization.

Challenges of a Private Cloud

Having said that, a private cloud is not the right solution for every organization.  There are a few constraints and challenges that would make the public cloud model more appealing to a lot of organizations.  There are inherent challenges with private cloud that need to be addressed before an organization can venture down the path of building one.

  1. Upfront Capital Cost: One of the drawbacks of private clouds is that organizations still need to buy, build and manage the cloud infrastructure, which defeats the primary premise of cloud computing.  One of the key value propositions of cloud computing is that it drastically reduces the upfront capital cost of in-house infrastructure, while providing the same or better service for a simple recurring operational cost.   This benefit cannot be realized with private cloud infrastructures.
  2. Time and Resources: Not all organizations have the time or resources with in-house expertise to build the infrastructure and automation required to stand up and operate a private cloud.  Case studies show that it involves much up-front investment in time and resources compared to simply going with a public cloud.  However, there are many up and coming startups that are offering private cloud related services and products that organizations can use in their internal data center.   While these products still need time to mature and become mainstream, this represents a significant shift towards the availability of the technology required to install and manage private cloud computing infrastructures.  These product and service offerings could mitigate some of the risks and challenges and blur the line between a private and public cloud.
  3. Size of the organization: Not all organizations can afford to build a private cloud for the two reasons listed above.  Crafting a business case for building a private cloud for a smaller organization is difficult, since building the in-house infrastructure for the private cloud does not provide as much of a return on investment as larger cloud deployments do. The public cloud provides impressive benefits related to economies of scale, and smaller organizations will find it difficult to build a private cloud solution that can match that.

So, what’s an organization to do?  The answer to that question depends on the size, resources and needs of the organization, and will be answered differently for every organization.   Private clouds are driven as much by what CIOs don’t like about public clouds – concerns about data security, corporate governance, and reliability will affect the decisions made by organizations. Private clouds theoretically can deliver some benefits of cloud computing without the pitfalls.   Additionally, cost will be a major factor to tilt the scale either ways — Wikibon argues in this excellent post that a private cloud is more effective for organizations over $1 billion.  As a result, it can also be argued that there is higher probability of organizations experimenting with private clouds prior to dipping their toes into the waters of public clouds.

Summary

Private cloud is not an oxymoron and is a perfectly valid model for the right organization.  Neither the public cloud or private cloud is an ideal solution for every organization.  For some organizations, a private cloud will be a stepping stone to a fully public model, and for others it will be the ideal solution. Organizations like the federal government with stringent data security needs or organizations with the right size/scale might pursue a private cloud model.  Availability, security, compliance etc. may drive other companies to look at a private cloud model as well.  In any case, cloud computing represents an exciting evolutionary step in computing and organizations will be able to choose between private, public or hybrid cloud offerings based on their unique needs and concerns.

Enhanced by Zemanta
(Visited 579 times, 1 visits today)