The Heterogeneous Environment of Clouds: Hybrid, Multi-Cloud and The Normalization Factor

Aug 6 2014 | by Boris Goldberg

40 Flares Twitter 0 Facebook 0 LinkedIn 40 Email -- Filament.io 40 Flares ×

When looking at the evolution of enterprise IT, it goes without saying that over the past few years, we have been well on our way to the cloud. Having worked as a development manager at a well-known software company, I dealt with my fair share of situations that could have been easily managed with the cloud. At times, additional computers were needed for specific short term projects, requiring the involvement of many bodies within and outside of the company to eventually finalize appropriate server purchases for the allocated task and budget. It is no surprise that enterprises have welcomed the cloud with open arms. Leveraging both the private and public clouds saves enterprises precious time and money. The private cloud enables enterprises to organize their repositories, while the public cloud provides supplementary resources.

The Enterprise Heterogeneous Environment

Should enterprises work with a single provider? No. I predict that the standard for working within the public cloud will eventually constitute at least two separate providers per enterprise. Competition is key here, seeing as enterprises do not want to be locked in to a single vendor. Utilizing multiple providers’ services better equips enterprises to fulfill any moment’s needs.

The same goes for private clouds. VMware has been an integral part of enterprise IT for over a decade. Nonetheless, with no better alternative to virtualization, enterprises are broadening their horizons, utilizing OpenStack, as well as other competitive initiatives, such as Cloud Foundry. With OpenStack’s seemingly never ending list of contributors and investors, the largest open source project we have ever seen, does raise a few concerns. While growth is generally seen as a positive outcome, control must be maintained. There is no doubt that open source projects have provided a great deal of innovation and ease to the world of IT. However, when dealing with large projects, such as OpenStack, the future is still unclear.

The question is, how can you see all of these heterogeneous resources on a single pane of glass?

Enterprises ought to utilize at least two providers in both the public and private clouds. This way, on one hand, a company can justify its investment in hardware and on the other hand, it will begin to reap the benefits of the hybrid and multi-provider approach. Analyzing effectiveness and the structure of different resources becomes difficult when various stacks or pillars are incorporated. There is still no good solutions today to run this hybrid cloud environment, which is a problem, seeing as it appears to be the future of every enterprise. Fortunately, several companies are currently trying to solve these issues. One of which is Cloudyn.

 

The Normalization of Clouds

If you want to view and analyze data from various cloud vendors on a single pane of glass, it has to be normalized in some way. Every cloud vendor has its own terminology for a single concept or product. What Amazon calls ‘types’, OpenStack calls ‘flavors’, and so on and so forth. What Cloudyn does is normalize the data from the different vendors into one, common language. Collecting data from different vendors and identifying the common features enables us to create a common language, map various compute resources, and analyze everything from usage and cost, to reporting, and so on.

The same goes for prices. There are so many different price models in the public cloud for various services. However, prices entail an additional complication, insofar as there are no public APIs for any of the public cloud vendors, even Amazon.

Public cloud prices are published, and can be changed and customized for different users. On the other hand, the cost of a virtual machine in the private cloud must be calculated using four variables: the cost of the virtual CPU, the cost of the RAM, the cost of NIC (Network Interface Card used to connect servers to distincted segments) and the cost of the disk. The overall private cloud cost is computed first by knowing how much the physical host costs, then adding the overhead expenditures of any data center (i.e. cooling, electricity, human resources, etc…).

Normalization is the starting point that leads to making comparisons. For example, estimating price differences between public and even private cloud vendors can only be performed when the prices of all are normalized into one common language. The only difference, as mentioned above, is that public cloud costs are published, and private cloud costs are set according to IT policies.

Final Words
Modern enterprise IT is naturally made up of a wide variety of components. One vendor will never be able to fulfill the needs of every individual enterprise, alone. As a result, the native pace of evolution involves innovative solutions, like Cloudyn, that are able to normalize the environment. This ability decreases modern IT’s inherent complexity without harming key capabilities, such as agility and flexibility.

Start Optimizing Your Cloud Deployment

Take a 20-day free trial, view your entire AWS deployment in one place and identify opportunities for greater cloud-efficiency. Gain full transparency, accountability and control.
start free trial

Login

Forgot Password?

No account yet? Register.

40 Flares Twitter 0 Facebook 0 LinkedIn 40 Email -- Filament.io 40 Flares ×