8 Linux Security Improvements In 8 Years

Digitization is accelerating the evolution of business processes, and the cloud is a key enabler in assisting companies to break their business processes into their core components. This is empowering companies to generate greater value from their data and applications. This can be seen in transportation with Uber, retail with Etsy, communications with WhatsApp, and wireless payments from Apple.

These changes leave incumbent companies in a quandary -- how do they leverage the massive investments they have made in technology while moving to meet nimble competitors in an increasingly digitized world? The problem is further complicated by the need to preserve visibility and control, especially in regulated environments. The answer is adoption of a hybrid architecture using both traditional IT and one or more clouds. The question for enterprises is not only how hybrid cloud architectures work today but how they will evolve as the cloud evolves. 

There is no standard, industry-accepted definition of hybrid cloud. Some definitions include connections between traditional systems and cloud systems as a form of hybridity (yes, that is a real word), and some do not. IBM defines hybrid as the connection of one or more clouds to traditional (i.e. non-cloud) systems or the connection of one or more clouds to other clouds. Based on that definition we can see four phases of hybrid cloud implementations corresponding to the evolution of cloud overall:

The days of enterprises viewing cloud primarily as a means of infrastructure extension have come and gone. While infrastructure remains critical, it is no longer the primary reason enterprises now adopt cloud. Today enterprises are building hybrid applications with some processing in their traditional environments and some processing in the cloud. This change has had an impact on hybrid cloud. During the infrastructure-centric days, hybrid cloud was mostly a matter of connectivity and basic integration. In today's application-centric world, hybridity must include connectivity and integration, as well as DevOps, systems management, data management, and security.

As the cloud progresses so will hybrid cloud architectures. IBM envisions a world where applications will be composed from a variety of Application Programming Interfaces (APIs) from various vendors. These APIs will be purchased from many different clouds. In the application-centric era, the focus is on a single traditional-to-cloud integration; in the API-era, that expands to connecting traditional systems to many clouds. IBM believes that much of this API-driven development may actually involve workloads from distant clouds delivered through extended containers, for execution against the customer's data in the customer's "home cloud." This new form of hybrid cloud will affect incumbent and new competitors alike.

[Want to learn more about IBM's plans for hybrid cloud? See IBM Promises Bridge From Data Center To Cloud.]

The key to hybrid cloud for this phase will be standard (either formal or de-facto) mechanisms for using containers to deliver workloads to customers. These containers will require extensions from today’s containers to add elements such as visibility, control, automation, and security. IBM expects that the universal hybrid cloud will be largely implemented via open standards and open source solutions.

The final phase of hybrid on IBM's visible horizon involves the data-centric cloud. While all the details of this phase are beyond the scope of this article, the matter of data-as-a-service can serve as an example. IBM believes that many companies will engage in the sale of data sets to other companies. The buyers of these data sets will run sophisticated analytics-based applications not only on their own private data but on data purchased from others as well.

The broad adoption of this business model will provide one of the harbingers of the data-centric cloud. As usual, the architecture of hybrid cloud will have to evolve to keep pace with the evolving focus of cloud overall. Data hybridity will involve careful consideration of data location. Even over today's high-speed networks, moving massive data sets is not easy. In the era of the data-centric cloud, customers will have to find ways to automatically select and ingest portions of large for-sale data sets. For example, an American property insurance company subscribing to real-time weather data would specify only US weather data from a weather vendor. The ability to make these kinds of selections and mash-up public and private data will be a key consideration of data hybridity during the data-centric era.

I should note that the phases are cumulative. The benefits of cloud infrastructure carry over to the application-centric phase since the applications are run on a cloud infrastructure. With companies operating in every phase today, the API-centric and data-centric phases are already underway among early adopters, and will become mainstream over the next few years.

[Did you miss any of the InformationWeek Conference in Las Vegas last month? Don't worry: We have you covered. Check out what our speakers had to say and see tweets from the show. Let's keep the conversation going.]