Cloud computing continues to be a major force and 2015 will undoubtedly be a year that sees it evolve. Here are four predictions for what will happen in the cloud market in the new year.
1. Hybrid cloud wins!
Arguments about the validity of public versus private clouds will diminish, as they did for the latter half of 2014. Hybrid cloud computing will become the norm driven by the following factors:
- IT’s increased comfort with all forms of clouds
- New specific hybrid cloud management tools
- Industry-wide recognition that the best execution cloud venues of applications, use cases and release cycles will vary greatly based on price and performance objectives, geography, data governance along with regulatory needs and requirements
2. For hybrid cloud to work, it will become all about the application
Whether they are simple or complex hybrid cloud environments, their inherent diversity will continue to become more complex as cloud service providers continue to add unique infrastructure and service capabilities in an effort to further differentiate themselves from others in their market.
This trajectory brings to light how very difficult it is to migrate, move and manage applications to and between these very different physical, private and public cloud environments. This complexity puts the spotlight on cloud migration and management platforms. Despite fairly rigid tool approaches in the past, cloud management platforms optimized for hybrid environments must be able to abstract above the complexity of the underlying cloud infrastructures and services by focusing solely on the application, its requirements and dependencies.
For hybrid cloud computing to scale, there must be two key attributes supported by any tool or process: portability and manageability. While seemingly overly simplistic, most approaches to cloud migration and management have so far forced an inconvenient choice of one attribute at the expense of the other.
Early script or orchestrator-based migration and management tools required the creation of custom scripts, cloud-specific VM images and modifications to all but the simplest application. This services-intensive approach delivered a migrated application with granular management capabilities but very limited portability. By “hard-wiring” the application to a specific cloud, very little of the initial migration effort is available to be re-purposed for any other cloud environment.
To address the issue of portability, image and container-based approaches have emerged. These approaches make it much simpler to grab one or many images or containers and lay them down on almost any cloud environment. These approaches however are not without their drawbacks.
Image-based approaches require movement of payloads such as operating systems and databases that are not needed. This approach tends to be expensive on the cloud, often introducing a V to V overhead of up to 27 percent, and most importantly, only deals with a small part of the problem. Most applications are made up of multiple VM’s, images or containers. These still need to be orchestrated, governed and managed — none of which is addressed by image-based migration approaches that deliver portability without any manageability.
Application-based cloud management platforms decouple an application, its topology, requirements and dependencies away from the underlying cloud. Rather than forcing an application to conform to each cloud, these tools get the clouds to dynamically provision their best practice resources on binding based on the application’s needs, laying down the app natively on each cloud and leveraging its unique capabilities.
This focus on the application — not infrastructure along with the maturation of benchmarking tools — provides insights into determining the best execution cloud venue for different workloads and will be crucial in enabling a truly efficient hybrid cloud-computing model.
3. Hype over containers will subside
It’s hard to imagine a technology getting more hype than containers did in 2014. Simply due to reversion to the mean, containers will not dominate conversations to the same degree in 2015.
However, the real threat containers pose is to the hypervisor. Running containers on top of virtual machines adds a layer of abstraction that unnecessarily adds overhead. Containers on top of bare metal, especially bare metal clouds, is a different story altogether and will be a legitimate threat to the hypervisor at some point.
The other potential aspect of container usage that might fuel a second hype cycle is the evolution of the community of migration and management vendors, which will undoubtedly fully support containers going forward. This adds even more value to the container story as containers become just another landing spot for software orchestrations that these platforms already support for installing software on virtual machines.
Other features of migration and management tools, like setting and enforcing governance policies and different levels of monitoring and management for different parts of an organization, will enhance the container management experience significantly.
4. Sophistication of hybrid cloud management platforms will shed light on hybrid hype by many CSPs causing them to truly shift to hybrid
Businesses have voted. No company wants to be locked in to any one cloud provider. Businesses recognize that hybrid will win the day with different applications belonging naturally to different on-premises and cloud environments. To align with this perspective, it seems that almost every cloud service provider (CSP) has raced forward with their messaging about their hybrid cloud capacities.
However, the realties of providers’ capabilities have fallen far short of the hype (#stopthehybridhype) with many covering little more than spinning up a VM on some other cloud, let alone the real need of managing applications fluidly between diverse physical, private and public clouds.
As new and very capable hybrid cloud management platforms become more mainstream and businesses realize that they can fulfill their vision of hybrid cloud without relying on a CSP, the CSPs — yes, even the largest — will have the rug pulled out on them. They will realize that they will need to either get aggressive, stay in control and drive true hybrid cloud … or they will be left out of control and on the diminishing side of being a spoke on someone else’s hybrid plans.
David Cope is executive vice president of corporate development at CliQr. With more than 25 years of executive leadership and marketing experience in fast-growth technology businesses, he is responsible for business development, strategic alliances, brand management and integrated marketing communications. Previously, he held executive positions at companies from startups to large publicly held companies. He received numerous marketing and technology-related awards and is an inventor on multiple technology patents. Follow him at www.linkedin.com/in/davidjcope.