"I've been reading a lot about "cloud computing" over the past few months and am more confused than ever. I'm not sure I understand what it could mean for me or my business, and how does it relate to virtualization? I thought that was part of the cloud."
You’re not alone. The term “cloud computing” has been used to describe a wide variety of services and frameworks recently, which has caused confusion among many in the industry.
One definition of the cloud is a service or services provided by an outside entity, such as an application stack you access from your systems, running on servers and storage maintained by the provider. In this definition, Google provides a cloud service such as Gmail or Google Docs that you might use but not need to maintain.
Another cloud definition is a lower-level service such as Amazon’s EC2. This service provides server and storage instances that you can then manipulate to run your own applications and services. Essentially, they build and maintain the servers and storage, and you are responsible for installing and maintaining the applications that run on those instances. In this particular case, the servers and storage are not physical, but rather virtual instances that are largely independent of the hardware. By removing the need to build and maintain your own hardware, services like EC2 can reduce the time and expense required to spin up new applications, and they can scale out faster and easier than traditional builds.
But there are even more types of clouds. The examples above are public clouds that can be utilized by anyone. A private cloud is designed and built in a similar manner, but is intended to be used solely by a private company. For instance, a company may have a need for server and storage instances to be readily available to internal working groups for any number of purposes, such as for software development. An easy way to provide for this business requirement is to create a private cloud that is available only to employees. The underlying framework of the private cloud is essentially the same, leveraging server and storage virtualization technologies, and is wrapped in a self-service portal tied to internal user authentication.
Thus, a development group that determines the need for a set of servers and storage to design and test new software would simply log into the private cloud interface and select the resources they need, and the back-end infrastructure automatically provisions and deploys the necessary components without requiring interaction by IT administrators. The end result is that IT has more time to build and maintain the infrastructure and the end-users get their resources faster and easier than the traditional model of sending requests to IT and waiting for hardware and software to be built and configured.
Of course, it’s possible to use public clouds to achieve the same results, but with a private cloud, all the data is maintained in-house, and thus the company has more control over where that data actually resides, versus the public cloud where you entrust the security and integrity of that data to the cloud provider.
While the concept of a private and public cloud might be similar, the way a company utilizes them is vastly different. A company that is heavily dependent on a public cloud provider will need plentiful and reliable Internet bandwidth in order to ensure that their access to their cloud-based resources are fast and stable. In some cases, this might even involve dedicated data circuits running from the company to the cloud provider, rather than standard Internet access.
In many cases, the use of VPN tunnels from the corporate infrastructure to the cloud infrastructure is also desirable in order to maintain the security of the data moving between the provider and the corporate network.
This type of cloud use relies on high-speed network edge switching and routing. When tapping into remote resources, features like QoS (Quality-of-service) and VPN termination of high-speed networks is a must, and solutions like the Cisco 3900-series routers can provide the tools necessary to make full use of public cloud-based services.
Building a private cloud requires the design and construction of internal switching and routing frameworks to support a large number of physical and virtual servers and storage devices. For the physical tasks, solutions like the Cisco 4900 or 6500-series datacenter switches make sense, as they can aggregate large numbers of high-speed data links into a single modular chassis, and optionally provide internal switching redundancy. Larger implementations might look towards the Cisco Nexus 3000, 5000, or even 7000-series to provide a more substantial backbone and provide 40- and 100-gigabit switching.
On the virtual side of the private cloud is the Cisco Nexus 1000V virtual switch. A virtual switch is a software-based network switch that couples with hypervisors running virtual server instances. In this way, the virtual switch can maintain network pathing, security, QoS and other elements of specific virtual servers no matter what physical host they happen to be running on at any given time. It’s essentially a network switch port that follows the server through the virtualized infrastructure, maintaining the proper network configuration.
The way you choose to leverage a cloud service is going to be highly dependent on the immediate and future business requirements, and you may find that a blend of the two is the right move, perhaps building a private cloud for internal use and leveraging a public cloud to provide services to your customers, with high-speed, secure connectivity between the two. Either way, cloud computing can make deploying new applications and services, fast, easy, and cost-effective.
This story, "Clarifying Cloud Service for Your Business" was originally published by BrandPost.