By Luke Norris, CEO and Founder, Faction
When organizations decide to deploy cloud services, the technology they choose matters tremendously. While pilot projects will tend to start small, with limited functionality, it’s important to choose a technology foundation that offers room for expansion—both in terms of functionality richness and hybrid cloud interoperability for the future.
"Microservices depend heavily on network-based API calls for communication between services—highly sensitive to latency between services"
The Efficacious Usage of Hybrid Cloud
Hybrid cloud is hot buzzword and a key tool in your cloud tool belt, but many organizations have little experience with coherently implementing it. Hybrid cloud is the deployment of computing resources across both public and private cloud infrastructure; some definitions extend that to include mixing physical gear in colocation into cloud environments.
Applications are at the heart of the drive to hybrid cloud, because the best hybrid cloud architectures blend applications running on a private cloud environment with the tools and options of hyperscalar public clouds. The public cloud appeal causes problems for some traditional IT workloads. The typical enterprise has not only their IT crown jewels – ERP systems, databases, CRM systems, and so on–but hundreds of line-of-business applications that deliver enormous value every day, but without the resources for developers to prepare them to run in the public cloud.
This creates a challenge for IT departments; how can you enable developers to utilize public cloud in areas where it makes sense, while preserving those other enterprise applications? Moreover, how do you enable these applications in the public cloud to access the other applications of the enterprise—because in today’s world, no application is an island.
This bifurcation between the need for stability in existing applications and agility in new applications led to a new term to describe the curation of those two halves of the IT world: Bi-Modal IT.
Bi-Modal IT: The Application-centric driver for Hybrid Cloud
Gartner coined the term “Bi-Modal IT” to describe an application deployment strategy where IT departments maintain a more traditional IT environment for their “Type 1” applications–those that require stability over agility and thrive in the traditional high-SLA, high-uptime world of traditional enterprise IT. The “Type 2” applications are those focused on delivering agility, and leverage Devops deployment models, cloud-native application development tools like NoSQL databases, and horizontal scaling models.
This fits perfectly into the first phase of a hybrid cloud model because existing Type 1 workloads can remain in a private cloud environment or transitioned to a hosted private cloud environment, typically VMware-based while providing a platform to blend those private cloud applications with public cloud resources. Hybrid cloud allows organizations to keep their people, processes, and technology to maintain their existing applications, while having a tight network integration to the public cloud, enabling interaction between private and public cloud applications. A significant driver for this is the shift towards new application architecture: microservices.
Microservices is a new paradigm for deploying applications that relies on each tier of an application being its own horizontally scalable layer. The servers in each tier typically have a single job, for example, an authentication service. With microservices, each tier scales out independently and the services become building blocks that can compose more complex services. Composability is a hallmark of cloud native applications, which are driven by application programming interfaces (APIs), even for intercommunication. Composability via APIs is the object-oriented programming of cloud native applications - it allows isolation, scalability, and re-use.
Because microservices depend heavily on network-based API calls for communication between services, the components are highly sensitive to latency between services. That typically means all the services need to be in a local or metro network area. This is a key factor driving the adoption of hybrid cloud: the need to maintain a private cloud environment, but a desire to have fast, low-latency connectivity to public cloud resources to take advantage of public cloud as well.
Hybrid Cloud to the Rescue
Hybrid cloud is meant to solve both the traditional IT environment and the next-gen agile application environment, and help them thrive in harmony. In a hybrid cloud environment, IT teams enable deployment of both traditional applications–using the people, processes, and technology they already know and rely on–as well as the “Type 2” applications, using that high-speed cloud-attached networking. This is a powerful combination, allowing the “lift and shift” of the traditional application environment, so it can be proximate to the new generation of application.
Most IT departments have a VMware-based environment that runs their traditional applications. They rely on this VMware layer to allow non-disruptive maintenance via vMotion, as well as the resiliency provided by VMware’s HA technology. Many leverage it for features such as snapshots or changed-block tracking in order to make backups and recovery easier. Some utilize security features to ensure the security of the guest virtual machines. Some utilize it for virtual desktop environments.
IT departments rely on their VMware skill sets to enable all this functionality. Consequently, many successful shifts to hybrid cloud look to utilize VMware-based environments that have connectivity to the public clouds.
There are three significant challenges to this shift, however:
• Many organizations do not have the experience, technology, or scale for simple or cost-effective connectivity into the public cloud from their private cloud environment
• Managing hardware, colocation, network gear, and so on means IT departments are stuck handling the world of capital budgets and hardware procurement cycles - but also have to deal with the billing, budgeting, and performance challenges of public clouds
• Many hosted private cloud environments force a shift on an IT team toward a managed services platform, even if they have the expertise to manage the environment, because hosted private cloud providers are often loathe to “co-manage” gear and provide true administrative access—which also interferes with their desire to sell additional managed services
The end result is that many IT departments don’t like either option: managing a hardware platform and datacenter space is time-consuming, costly, and doesn’t fit with the agility and economics of cloud adoption; but they also aren’t keen on throwing away the people, process, and technology they rely on for their “Type 1” traditional workloads.
Do you get a cloud for your applications, or fit your applications to the cloud?
This dilemma is resulting in a new generation of hybrid cloud options being rolled out by forward-thinking providers. Discerning IT departments are demanding hybrid cloud environments that allow them to continue to use their people, processes, and tools to manage their private cloud environment, while simultaneously having tight integration to public cloud environments. Some hosted private cloud offerings now come with turnkey public cloud connectivity options; other IT shops may turn to a network provider or datacenter provider to help solve that problem.
Either way, IT professionals are realizing that having the best of both worlds maximizes their agility and minimizes their cost, and we can expect demand for hybrid cloud to increase as microservices-based architectures become more prevalent.