Cloud Computing

Ξ Deixe um comentário

The inevitability of Hybrid Cloud

publicado por Abilio Bueno de Oliveira

Figura - The inevitability of Hybrid CloudWho could have ever imagined that born-on-theweb companies like Uber, TripAdivsor or HomeAway would challenge established companies and industries segments forcing traditional business modeling to transform their way of doing business in order to survive in the market?

As stated in my last post (From Mathematics to Cognitive Computing), the world of computing has entered the next stage of evolution, the Cognitive era, and that happened while we were debating the best execution venue for the applications, whether my infrastructure remains “on premise” or not.

Many specialists say that the future belongs to Public Cloud which Hardware and Software resources can be available in affordable cost, in the other hand another group defend the trending of going to a Private Cloud due some information security or confidentiality requirements.

So again, for each of the Business lines or Systems that are supported by the organizations, we will have advantages or disadvantages in using whichever model we want to use. To understand which model fits to a certain reality is important to collect and analyze all business and technical requirements in order to make an assertive decision.

I am not going to put an end to this discussion, but I venture to say that there is another option that can bring the advantages of both models.

This option is called Hybrid Cloud.

An as I stated in the title of this text, Hybrid Cloud is inevitable!

Ok, that is not new, but why is it inevitable?

The new generation of dynamic and customer focused applications requires different needs.

The data generated by these new applications are highly volatile which drive us to not maintaining that in a high cost infrastructure, but the output of this process is a valuable knowledge that needs to be stored in a safe place.

Imagine an analytics workload may require compute-intensive statistical processing with complex queries manipulating 30 Petabytes (PB) of data and a large amount of CPU, but the after mining this data; we get into 2 Terabytes (TB) of knowledge that is really useful. However, the analytics workload may only require resources once a quarter when analysis is conducted, for example.

It may be a clear decision that you don’t need to have “On premise” the 30 PB neither the Compute Platform; it will be expensive and not aligned with the business needs. But the 2 TB of output we may decide to keep it close to us.

One of the biggest challenges for organizations is ensuring that it will be possible to create predictability when managing different workloads across a huge variety of environments.

In a Hybrid Cloud solution the idea is provide the capability to dynamically change as deployment models change and adjust the environment to new workloads and new services.

The clear benefit here is the flexibility we gain to select the right environment for the workload.

That is possible because a hybrid cloud model are built in Open Standards as Linux, OpenStack and Hadoop, so the systems can “talk to each other” even if they are from different vendors or if they are in different places, providing interoperability between on premise systems with systems in Cloud.

Different workloads have different characteristics, and the platform for them to run in an optimized manner will also differ.

For example, a transactional workload may necessitate round‐the‐clock availability and high speed but only require a small amount of compute power.

Matching workloads with the right environment allows organizations to optimize efficiency while also only budgeting for the resources that are required.

Another benefit of a hybrid cloud environment is to have the capability to use the best deployment model in terms of cost that offers the level of service demanded by your customers.

Therefore, at times a public cloud service is ideal (testing, prototyping, managing huge volumes of unstructured data that are analyzed on a one‐time basis).

In other situations, a private cloud may be the best choice, such as when data is too sensitive to security or compliance to risk managing it in a public cloud. In addition, if the application becomes a profit center, it may make sense to keep the application in a private cloud to avoid cost overruns.

How to move on?

Without changing the way that we think about IT, this transition can’t happen.

I may suggest that you study the pattern of actual workloads, review the reference architecture, collect all the non-functional requirements and see how it is possible to build an efficient infrastructure. Theses patterns can be used to assist you in understanding and addressing a particular business problem and how it can be solved via a Hybrid cloud solution.

Organizations that are transforming themselves are moving away from thinking about IT as a backroom activity. These companies are putting IT in the forefront of empowering innovation and transformation.

In other words, it means leveraging all the IT resources across the organization creating a flexible and scalable platform for innovation and change. This level of change requires a drastic reinvention of IT.

A hybrid cloud strategy that leverages the assets and computing models that exist with a continuum of emerging cloud models is at the heart of this transformation, that’s why it is INEVITABLE!


My name is Abilio Bueno de Oliveira Neto and I work as a Client Technical Client in IBM Asia Pacific. I work to combine technical knowledge with Business skills to provide advice and support on client Infrastructure. I have more than 22 years of IT experience. I live in Auckland - New Zealand.

Abilio Bueno de Oliveira


You must be logged in to post a comment.





Newsletter: Inscreva-se

Para se inscrever em nossa newsletter preencha o formulário.

Artigos Recentes