An Interoperability Framework for Enterprise Applications in Cloud Environments

Introduction

A cloud can simply be defined as a remote platform supporting the deployment and use of computer-based resources and services, in a setting characterized by elastic, dynamic and automated resource provisioning, paid as used and managed in a self-service way (Armbrust et al., 2010).

Through virtualization, a pool of physical resources (servers, storage, networks, and so on) supports the dynamic allocation, provisioning, decommissioning and release of virtual resources, forming an apparently elastic fabric of resources that are used on demand and paid as used. However, users want the services that resources support, not the resources themselves. Resources are increasingly seen as a commodity (Carr, 2004), allowing some IT enterprises, the providers, to specialize in providing IT-based services and resource infrastructures cheaper, more reliably, better managed, faster provisioned and in a more scalable way than any of the organizations that just require these services and resources for their business activities, the consumers.

This dichotomy between consumers and providers, a marriage of convenience as any outsourcing agreement, is known as utility computing (Brynjolfsson, Hofmann & Jordan, 2010). Providers take care of many issues on behalf of the consumers, namely expertise, risks, costs and management, in what concerns resources and generic services. Consumers still have to tackle a part of the overall solution, namely application-specific services, but they can concentrate more in their core business and less on the IT technologies that support it.

After an initial period of slow growth, in which concerns about security, privacy, performance and availability acted as inhibiting factors, cloud computing is finally blossoming. For consumers, it is now easier, cheaper, and faster to get computing resources from a cloud than by resorting to conventional IT systems and applications. In most cases, the advantages now outweigh the risks and other disadvantages. All large IT providers are now investing heavily in cloud computing.

This is clearly a market driven by providers, with consumers still cautious about the transition, but the scenario is evolving at a fast pace, both for individual and enterprise consumers. Gartner analysts forecast that the public cloud services market will grow at a compound annual growth rate (CAGR) of 17.3% over the 2014-2018 five-year period (Anderson et al, 2014, May 13). Gartner also expects that end-user spending on public cloud services will grow 19% in 2014, up to $158 billion, with annual growth rates of 45% and 33% for infrastructure (IaaS) and platform (PaaS) cloud services, respectively (Anderson et al, 2014, March 31). These rates clearly indicate a shift in spending, from traditional IT systems to cloud services.

Two relevant factors gave their contribution to the bootstrap and boom of cloud computing:

• •

  For individual consumers, social networking and multiplatform mobility (laptops, tablets and smartphones) raised the need to store information in a server somewhere, always available to be accessed seamlessly and in a synchronized way across platforms;

• •

  For enterprise consumers, the market pressure caused by increasing global competition and ever-shortening turnaround times, combined with a sluggish global economy, emphasized the basic principle of concentrating on core business and (dynamically) outsourcing the rest.

However, the problem that drove the appearance of the Web, interoperability, remains unsolved. The goal is to endow distributed systems with the ability of meaningfully exchanging information in interaction patterns known as choreographies. The problem, unfortunately, is even worse today than 25 years ago: