March 14, 2015 | Written by: IBM Cloud Staff
Share this post:
By Michael Skott and Hans-Jürgen Kunde
Infrastructure as a service (IaaS) is used in many of today’s cloud environments, but alone it may not be enough for enterprise cloud solutions.
In his post “What is infrastructure as a service?” Michael J. Fork provides a number of potential definitions of the term. In our own discussions with enterprise clients seeking to establish data center and cloud solutions, we have found that even when the definition is agreed to, IaaS cannot be used in isolation if you plan to run productive enterprise workloads.
This post aims to help you find the right approach for you, and to illustrate why it is important to save time and effort with application automation on top of the infrastructure.
We can break down IaaS into three distinct dimensions, each building on the previous.
The first dimension of enterprise workloads covers the component layers of a typical application server. The following image provides a comparison between a traditional information technology (IT) infrastructure and IaaS.
A typical IaaS provides you with everything up to and including the operating system (OS) installation, as you can see from the items in blue boxes above. Compared to traditional IT, IaaS does not include the management of the operating system, installation of middleware and databases, runtime installations like Java engines or other programming runtimes, or application and custom data. This IaaS model is typically used for simple development and test workloads, or unmanaged client application services.
Having a number of server images provisioned as IaaS gives you a number of standard installed operating systems, but it doesn’t yet provide business value. Business value is added through middleware, database and application services on top of the managed OS that forms the enterprise application.
To achieve business value, there is more to be done:
- Additional components need to be installed and managed on top of the IaaS.
- Compliance requirements are to be fulfilled for security and data privacy.
- Availability and disaster recovery have to be considered.
This drives the need for the second dimension of an enterprise workload: the service management across the vertical layers of the stack, incorporating the different service providers. This is where information technology information library (ITIL) concepts and best practices come into play. Cloud IaaS does not change the principles here; it only changes some automation aspects.
The third dimension comes into play if you leave a single-premises scenario and connect hybrid cloud elements by running workloads on different cloud environments depending on workload class and provider. From an enterprise management perspective, you still want to know which workload is running where, under what circumstances. You also need to define a governance framework across different environments or providers to establish the rules under which a workload can be placed somewhere. Many organizations underestimate the amount of effort this can take, and this misjudgment can cause complications after the fact. Aspects of service integration (such as the service integrator role) and brokerage services may come into play.
IaaS provides flexible pricing and fast provisioning of the basic infrastructure needed to run an OS. Within a couple of hours, you can spin up as many operating systems you like. What’s even more interesting, though, is that you can automate several processes, including the installation of the additional components mentioned above, the setup of the required level of service management and the choice of a government framework that suits your enterprise.
If you feel these dimensions are of interest for you, tell us about your experience and please post your comments below or contact us on Twitter @michaelskoett and @hjktweets.