Despite the vast quantity of talk and enthusiasm for all things cloud, few CIOs and IT architects see the future of their analytics infrastructure as an either/or decision. A more apt description for their designs is hybrid, a word that most commonly means a computing environment that is partially either on-premises or in a private cloud, as well as to some extent located in the public cloud.
One of the challenges for those who need to evaluate hybrid solutions is that the cloud has become a catch-all term that, through overuse and generalization, has taken on a deceptively reassuring air of simplicity. That shorthand masks what is actually a highly textured landscape that forces one to undertake quite complex considerations. How many really understand the differences between a managed cloud, public cloud, private cloud and hybrid cloud?
Words are being used in ways that obscure the true nature of how computing resources do or do not work together. More significant is that oversimplification masks the potential benefits that can be realized
Among the primary sources of confusion is the question of just what hybrid means for data and analytics. The mere fact that a solution can run either on premises or in a cloud setting does not make it hybrid in my view. It’s flexible in its deployment options, but it is not a hybrid solution.
Here’s my definition of a hybrid cloud solution: A hybrid cloud solution is a mix of at least two technologies (private, public, managed or on premises) that are orchestrated to work together.
If the data platforms that are on premises and in the cloud can’t talk to each other or, worse, don’t even know that the other one exists, those are separate silos, not a solution. Without that fundamental capability, there’s really not much point in being “hybrid.”
A true hybrid architecture is deployed both in the cloud and on premises, and, moreover, the two (or more) systems are orchestrated so that at a macro level, all the components work together as a single solution. The systems are aware of each other, can actively leverage each other to address the requirements of particular analytics use cases, and ideally, benefit each other. A hybrid exemplifies the “whole is greater than the sum of its parts” philosophy.
A true hybrid solution enables an organization to leverage opportunities such as:
- Cloud-bursting to rapidly add capacity or shift on premises workloads to the cloud to optimize compute processing. That’s what will keep meaningful insight flowing, whether to support seasonal spikes or prepare for a busy Monday morning strategy huddle.
- Spinning up cloud data labs quickly to experiment and explore analytics techniques.
- Cloud disaster recovery that keeps data-driven operations online by automating failover between different instances.
- Data sharing with disparate applications or third parties by granting access to warehouse resources.
In order for an organization to realize such benefits, there needs to be orchestration between different instances. The system needs to dynamically route users and support data and object synchronization across the hybrid environment. It should support dynamic cross-platform querying. In other words, the plumbing needs to be in place. When new data comes in, awareness of that data needs to be synchronized across the deployment. A query created in the cloud or on premises requires a system that is aware of the entire hybrid environment in order to effectively deliver meaningful results.
The problem to date is that most solutions run on-premises and in the cloud; they are not integrated to support a hybrid approach in which the whole is greater than the sum of its parts. If one department is using technology in the cloud and somebody else is using it on premises, that’s not a hybrid solution and you’re not fully realizing the benefits of a hybrid cloud architecture.
Given the critical nature of analytics, we need to be able to run queries across platforms and leverage all the data that we have in service of the business. A hybrid cloud solution is in fact what IT decision-makers are asking for. Getting to know the different hybrid cloud models, their value and their potential return is a good first step toward identifying a sufficiently flexible solution to keep your organization’s options, and opportunities, open.
Chris Twogood is Vice President of Product and Services Marketing for Teradata Corporation. He is responsible for marketing Teradata products (database, utilities, and platform) and Teradata services (professional and customer services), plus the technical field sales support teams. Chris has 25 years of experience. Chris has extensive experience in the computer industry specializing in data warehousing, decision support, customer management, and appliance platforms. He started his career with NCR in retail point-of-sale (POS) solutions and then moved to AT&T to manage channel development and strategic partners. Then Chris joined Teradata. At Teradata, he has had roles that span strategy, application definition, marketing, product requirements/management, platform solutions, and product marketing. Chris holds a Bachelor of Science degree from California State University at Long Beach with an emphasis in marketing.