[sdm_download id=”6948″ fancy=”0″]




At the recent IBM Cloud Analyst Summit, one message was clear: The centerpiece of IBM’s cloud strategy is hybrid cloud systems. IBM executives discussed the company’s strategy of delivering a “spectrum” of cloud and on-premises solutions to meet the needs of its many different types of customers. To manage the complexity of hybrid cloud environments, IBM is using a combination of open source solutions within its IBM Bluemix DevOps[1] platform and a new “hybrid cloud controller” products.

Hybrid Clouds are Complex

At present, IBM offers customers a full range of cloud services, from infrastructure as a service (IaaS), platform as a service (PaaS), and software as a service (SaaS). IBM is delivering these solutions as multi-tenant cloud instances, as private instances within the IBM SoftLayer cloud, and, of course, supports on-premise solutions –as well as a combination of all of these options.

Hybrid cloud environments combine cloud and non-cloud assets such that they work together as a single, end-to-end system. Applications are deployed onto whichever type of infrastructure or platform best meets the needs of the application and the organization. Often, customer concerns about data security, and compliance with governmental regulations, result in some cloud solutions being on-site, inside the enterprise data center – while other elements of the hybrid solutions reside in public clouds.

For example, if an application needs a high performance environment, it can be moved into a private cloud instance with dedicated servers and storage. Another application may need to be highly secure — and only an on-premises deployment provides an appropriate level of protection. Common components and applications with less stringent requirements may be perfectly fine running on multi-tenant cloud instances.

IBM has, for a long time, been customer-centric. It is IBM’s strategy to offer any combination of cloud or non-cloud hardware and software resources that a customer may require. Customers and applications, in the IBM view, exist along a spectrum of resource needs and IBM plans to supply as much as it can. If needed, IBM also turns to partners, such as system integrators that can bring unique technology or services to bear on customer problems.


Hybrid clouds contain both private cloud instances and multi-tenant public instances, as well as on-premises servers, infrastructure components, and microservices[2] . They are attractive to IT professionals because they allow the greatest flexibility in system design while maintaining proper performance, protection, and information governance. Individual software can run in the environment that is most suited to the needs of the software without constraints imposed from other applications. In many cases, hybrid cloud environments provide the best balance between cost, convenience and security.

IBM is not alone in promoting the build out of hybrid clouds. Competitors such as Dell, EMC, HP, and Oracle have also described their approaches to hybrid clouds – and the technologies they will leverage to build hybrid clouds on behalf of customers. The reason this competition is gaining in intensity is that customers are finally ready to phase in hybrid cloud infrastructure – often using technologies such as OpenStack cloud software to achieve it.

For example, within a hybrid systems architecture a credit card processing service can run in an on-premises, highly secure environment. End-user facing applications can be deployed to less costly and easily scalable multi-tenant cloud environments. The end-user applications can then access the credit card service when it needs to through secure, RESTful APIs. Each type of software runs on the most suitable platform and yet operates as an entire system.

Complexity Inhibits Hybrid Architectures

There are challenges in building hybrid clouds –and this is where the opportunity for vendors and system integrators is apparent. Hybrid cloud systems architectures are highly complex. Each cloud and on-premises environment is different, housing dissimilar system stacks, or versions of stacks, and having divergent performance, security, and accessibility profiles.

Deploying and managing applications and services across so many diverse platforms is a chore, to say the least. Migrating from one environment to another may require fine-tuning or a complete refactoring of complex software to accommodate different virtual machines, libraries, and hardware capabilities. Managing applications comprised of hundreds of small microservices spread over many cloud and on-premises instances is equally complex. In many cases, multiple instances of the same microservice are running on different cloud instances or non-cloud servers, which makes troubleshooting difficult.

And yet standards are emerging that will make it somewhat easier to “abstract” the microservices from the underlying infrastructure. These standards, emerging for software-defined infrastructure (SDI), focus on managing virtual machines (VMs) and software “containers” (containing many VMs) as they move across data-centers and in-between multiple clouds.

IBM Tackles Deployment in Hybrid Environments

IBM is tackling this complexity in two ways. First, IBM  offers a number of open systems deployment tools, through the IBM Bluemix DevOps framework. Currently, IBM supports Cloud Foundry for deploying PaaS software stacks and OpenStack for provisioning cloud infrastructure components. Together, these open source frameworks facilitate the deployment of software components and applications on a wide range of cloud and bare metal resources.

IBM also supports deployment of Docker containers[3] – IBM is a founding member of the Open Container[4] project – and traditional virtual machines (VMs) as well. By offering a variety of open deployment technologies and methodologies, IBM is providing a full and flexible toolkit for IT professionals trying to deploy applications to complicated hybrid environments.

The Hybrid Cloud Controller

Deploying to a hybrid cloud environment is one challenge. Managing it is another. IBM’s response to managing hybrid clouds has resulted in the announcement of what IBM termed a “hybrid cloud controller.” Not yet a product, the hybrid cloud controller (which is unlikely to be its real production name) addresses the problems of visibility, control, and – most especially – automation in hybrid cloud environments with potentially millions of entry points. The hybrid controller is envisioned as a way of automating not only deployment, but also governance and management within large-scale hybrid microservices architectures.


Every large software and systems vendor has begun to ramp up cloud services beyond the relatively simple initial product offerings. As was evident from Oracle’s recent announcements[5], the ability to extend cloud environments into higher performance computing environments is crucial to satisfying the needs of the most large enterprise customers.

[1] DevOps is derived from the terms development and operations. It is a methodology by which development and operations work together to create a seamless coding, testing, deployment, and monitoring experience for technical teams.

[2] Microservices are small independent software processes, accessed through language independent APIs, and assembled into applications functions.

[3] For more discussion on containers consult “Containers: The Next Wave in Application Deployment”, August 2014, Tom Petrocelli https://www.neuralytix.com/blog/2014/08/18/4838/

[4] The Open Container project is a multi-vendor project under the Linux Foundation.

[5] See the recent Neuralytix research note, Oracle expands IaaS and PaaS cloud services, June 2015, Tom Petrocelli and Jean Bozman fore more information. https://www.neuralytix.com/blog/2015/06/25/oracle-expands-iaas-and-paas-cloud-services/



[sdm_download id=”6948″ fancy=”0″]