Cloud Computing and Software-as-a-Service

Cloud computing has the potential to deliver a number of very important benefits, including the ability to provision compute and storage resources on-demand. These benefits take their most visible form in public clouds. However, for a variety of reasons, some applications and workloads need to be run from behind the firewall.

For example, the size of some data sets makes it impractical to transport over the wire to a public cloud data center. Management may have security concerns about sensitive data being processed and archived in a facility outside of IT’s control. Sometimes there are specialized hardware, architecture, networking, or storage requirements that cannot be accommodated in available public cloud environments. In response, many companies are exploring private clouds.

Once a private cloud is being considered, there are two basic approaches to deployment: Build your own cloud or purchase a cloud appliance.

Build Your Own Private Cloud

With most enterprises already running their own compute, storage and network resources, one approach to consider is redeploying these existing resources into a private cloud. In fact, given the last decade’s trend toward server consolidation, many of these servers may already be running a virtualization layer. Starting from this layer to build your own cloud, deploying infrastructure software (IBM, VMWare, etc.) is a logical consideration.

However, building a private cloud involves more than adding software layers to existing resources. “Today, enterprises and service providers that are interested in launching cloud computing services face the difficult task of integrating complex software and hardware components from multiple vendors. The resulting system could end up being expensive to build and hard to operate, minimizing the original motives and benefits of moving to this new model,” said Chad Collins, founder and CEO of Nubifer.com.

Unfortunately, many enterprises may not have the internal resources and expertise to take on this integration workload. This is where a consulting firm like Nubifer can play an integral role in solving these vexing problems.

The Open Source Alternative

In addition, with proprietary technology comes the risk of vendor lock-in. In response, open-source alternatives have emerged. Rackspace CEO Lew Moorman said his company decided to use OpenStack to open-source the software behind the cloud computing stack “because we believe a widely adopted, open platform will drive standards.” As of OpenStack’s six-month anniversary, 53 companies had joined the community.

Drawbacks to adopting open source do exist, however. For instance, OpenStack code base is still relatively young, and features such as supporting VMware hypervisors and live migration of instances are still in development.

In addition, IT personnel still need to download the project releases and integrate with the existing compute, storage and networking infrastructure. Although complexity can be removed by an open-source project, integration complexity still exists as components to the private cloud are provided by multiple hardware and software vendors.

This brings up another potential drawback: do you burden your internal IT staff with these modifications? Nubifer is here to help…

Cloud Appliances

Another approach to deploying a private cloud is through a cloud appliance. A cloud appliance is a rack of equipment delivered fully assembled and tested, with the software stack loaded and configured. When the appliance is connected to the power and network and turned on, it works.

It’s equivalent to buying a car today. Yes, you could assemble a “new” car from components already lying around in your garage or pieces from existing cars. Some people relish this idea and no doubt could come up with quite a vehicle. However, for the mass market, driving a car already pre-assembled and tested right off the lot is the preferred option.

As in the car example, a cloud appliance hides much of the hardware and software complexity required in integration and assembly. Instead, the appliance provider takes on this workload. And through replication and standardization, this provider has the opportunity to travel up the learning curve. It’s an equivalent argument that supports the move to Software-as-a-Service. The team that does the same thing over and over again — while developing supporting processes, test vectors, and underlying knowledge — will be more efficient than another “do-it-yourself” team that’s building an equivalent cloud for the first time. The result is these efficiencies translate into reduced overall deployment costs.

For example, computing solution provider IBM delivers a private cloud appliance. This appliance integrates standard hardware components and x86-based servers. By deploying a pre-integrated private cloud appliance, an IT team is saved time building its own. This enables an enterprise to focus on delivering its core business value rather than building IT infrastructure.

In addition, the IBM private cloud appliance natively includes private cloud management capabilities from Platform Computing as part of its cloud software stack.

IBM’s private cloud management offeringis an integrated solution combining self-service, orchestration, and automation for heterogeneous resource pools.

Getting back to the support question, a cloud appliance also offers a potential solution. The IT customer has only “one throat to choke” as the appliance vendor takes on the responsibility of supporting all the components utilized in building the appliance.

Cloud appliances have potential drawbacks. For instance, by definition new equipment is being purchased as part of the appliance, versus redeploying existing resources. For this reason, a company might consider an appliance during a hardware refresh cycle. In addition, as an appliance there are a limited number of pre-configured models, and one size does not necessarily fit all workload requirements.

An End to Parting Ways

During the early 1900s, there were hundreds of car manufacturers in the United States alone. By 1913, Henry Ford and his team developed the basic techniques of the assembly line and mass production, bringing standardization to the building of cars. The rest, as they say, is history.

Fast-forward to the present day evolution to cloud computing, and parallels can be drawn. It’s clear that companies are attempting to focus on core competencies, which for the most does not include building out IT infrastructure. And public cloud vendors are taking advantage of standardization to reduce costs and offer higher levels of agility.

However, many workload requirements inhibit moving data sets to public cloud environments, spawning the deployment of private clouds. However, when an enterprise considers building a private cloud, it’s back in the discussion of building out IT infrastructure.

Cloud appliances offer a potential solution. By pre-integrating all components, IT simply plugs in and turns the power on. And after all, when buying a new car, you would prefer to turn the key and go, versus huddling hour upon hour reading the user manual. Why shouldn’t your private cloud deliver a similar experience?

For more information on private cloud implementation contact a Nubifer representative.