Cloud Series: Definitions Pt 2



This is part two of the Cloud Series. See Part One here.

Part one took a look at “the Cloud” and drew a distinction between the cloud and cloud computing.

The cloud is just a service that is delivered over a network (oh so many of them now). Obviously, all these cloud services run on computers, so one might erroneously conclude this is cloud computing (clearly marketing folks seem to think so).

Where a cloud service describes a service delivered over a network, Cloud Computing describes a way to create a service. Cloud services may or may not utilize cloud computing.

Cloud computing is a fundamentally different approach and architecture to the creation of servers and data center services. This is not to be confused with colos. A colo is a professional data center that rents its space. Consider most businesses rent all their space – and a colo is just another lease that comes with HVAC and a few other services included in the rent. The servers are generally purchased and maintained the same way (as before) regardless of where they reside.

Traditionally, to create a service required lots of hardware. In 1999, during the Dot.Com Boom, every high tech Internet startup was investing in and building their data center(s). Data center’s require lots of care and feeding- redundant servers and networks, HVAC, security, administrators, etc. These are generally distractions from whatever the core business really is; but until recently, it was considered a requirement of doing business. (As a proof, today’s new dot.coms are not creating data centers).

Cloud computing separates specific core components such as CPU processing or storage and redefines them as abstract concepts. Basically, object oriented hardware. These core components became services themselves. The key technologies enabling cloud computing for voice services are virtualization, web interfaces, SIP trunks, and VoIP. Rather than pick a server for service, which picks a CPU capacity, a virtualized environment can provide CPU services as needed.

Cloud computing is often described as a utility model. Consider that lamp beside your couch. The more it is used, the more electricity it consumes. Cloud computing allows for the same concept with computing resources; effectively increasing the supply of CPU and storage as they are needed. Lamp designers and application developers are free to focus on what they do best. This is a significant change. Amazon’s cloud services requires no upfront investments or even a contract – only a credit card to get started.

A cloud service does not necessarily equate to cloud computing. A cloud service could be a PC in the basement (many are). Nor does cloud computing require a cloud at all – many of these technologies can be purchased and implemented in a private data center for internal services – this is called a Private Cloud.

Conversely, cloud computing services can be purchased (rented?) from several companies including Amazon and Google (they learned a lot about cloud computing as a prerequisite of their successes) with innovative pay as you go/pay as you grow financial models. This is referred to as the Public Cloud. These services can include virtual operating systems (IaaS or Infrastructure as a Service) for implementing purchased applications; or virtual environments ready for application development (PaaS or Platform as a Service). These providers tend to charge for service usage (typically CPU cycles, storage, and IO per month) at attractive rates compared to server ownership.

As usual, for those that don’t fit in either category; a Hybrid cloud could be created with elements of both public and private clouds. Potentially separating key services (or clouds) with some in the public domain and some in the private data center.
Upcoming posts in this series will deal with why cloud computing (and why the hype) now, and specific likely impacts to telecom/voice services.

Dave Michels