Strategies for Multi-cloud Management – Interview with Jay Litkey, President and Founder of Embotics
Is there a type of organization best suited for hybrid cloud adoption?
JL: In order to adopt hybrid cloud, organizations should have already reached a level of standardization and optimization that allows them to govern and automate their technology process. Virtualization and its successor, the “software-defined data center,” bring a level of operational flexibility to an organization, where applications and their workloads are managed first and foremost, and physical infrastructure becomes resources deployed in service of these applications. Hybrid clouds are a logical extension of this philosophy.
The speed at which organizations are able to adopt, however, is directly related to their existing team and process. We’ve found that the most successful organizations are those with the ability to look critically at existing processes, as well as organizational structure, and iteratively adopt new techniques and measure the results with the goal of achieving continuous improvement.
What are some of the challenges – from both a technology and organizational standpoint – associated with creating a hybrid cloud that IT organizations should be aware of?
JL: There are tremendous benefits to creating a hybrid cloud, including the control and security of private clouds and the operational flexibility, elasticity and lower cost of public cloud offerings. However, marrying the two can be challenging and requires careful attention to IT organizations, processes and procedures.
Moving to the cloud – regardless of whether it’s public, private or hybrid – reduces the physical number of servers, but it typically increases the number of server instances. This requires admin time, as well as systems to manage the entire services stack (operating systems, applications, network and storage).
While many public cloud vendors have effective management consoles for single hypervisor virtualized environments, there is a gap when trying to manage today’s multi-hypervisor environments and legacy hardware. The introduction of hybrid cloud infrastructures complicates things further, as most organizations have found that their current tools are not able to provide the vision needed to operate across environments effectively. Unfortunately, throwing more bodies at the task is not a practical or viable solution.
To get to the hybrid cloud, IT organizations will not only have to reinvent themselves (aligning security, operations, application, server, storage and network teams toward common goals), but also implement new integrated data center management systems that are uniquely designed for virtual environments. And they need to do all this while already bogged down in their day-to-day administration and cross-departmental coordination and education. Certainly not an easy task without the right team, technology and plan in place.
Do IT organizations’ processes and procedures need to change when moving to a hybrid cloud environment?
JL: It depends on the details of the organization and their current processes. If an organization has embraced the “software-defined data center” (SDDC) such that they have processes in place that manage applications and workloads first; where they abstract and aggregate physical infrastructure resources and apply them to these workloads as needed, then the move to a hybrid cloud environment is simpler and culturally compatible. On the other hand, if the organization is structured more traditionally and manages along physical infrastructure boundaries, then it will be a more challenging and comprehensive change to move toward a hybrid cloud environment.
What are the drivers of change you most commonly witness within IT organizations moving to a hybrid cloud?
JL: The drivers we most commonly see for hybrid cloud are varied. Most existing businesses have on-premise IT in some fashion. But cloud is pervasive and in the conversation. Many organizations also discover that they have “shadow-IT” going on organically, where business units have found IT to be non-responsive and have hence gone around them with public cloud deployments, but later on want IT to manage it. Other times, there is high-level executive pressure on IT’s budget influenced by the widespread perception of public cloud cost advantages.
These are some of the reasons organizations want to embrace hybrid clouds. Success though requires:
- Vision of the future. In this case, IT organizations have a vision of a hybrid cloud infrastructure and details of what type of workloads will run where.
- Dissatisfaction with status quo. We see this tend to occur as virtual stall starts to hit and initiatives slow down.
- Engaging your IT customers. Many organizations are embracing agile, DevOps style software development. IT needs to be embracing this style of service delivery. Be part of the solution, not part of the problem.
- Being agile and having the ability to iterate. IT organizations don’t need the whole plan in order to move in the right direction – just the first few steps. The additional steps will become clearer once you’re on the path.
Are there steps you recommend IT organizations take to help them adopt a more pragmatic approach to creating and managing hybrid clouds?
JL: Many organizations suffer from “paralysis by analysis” when it comes to implementing a multi-cloud infrastructure. Others are reactive, having pressure from “shadow-IT” using public clouds on their own. So much needs to get done and so many internal changes need to be made that, oftentimes, nothing actually gets done. Successful implementations tend to follow a more pragmatic, iterative approach.
Making sure your house is in order is imperative when creating and managing hybrid clouds. We’ve heard from leading analysts at firms such as Forrester and Gartner, and have seen this be the case within our own customers’ organizations, that a common set of virtualization management and cloud automation features are needed to create and run IT-as-a-Service (ITaaS) infrastructures, including: resource optimization, lifecycle management, workflow and automation, IT costing and chargeback, and self-service and service catalog. While all of these features are eventually required, they don’t all have to be implemented at once or comprehensively. In fact, it makes much more sense to start with small measurable improvements; look at implementing them as you need them.
We suggest breaking the move to a hybrid cloud into smaller, less intimidating steps that can be dealt with one at a time with the amount of focus required. If an on-premise, private cloud is your starting point, look for frequently provisioned, short-lived, sporadic or transient workloads as the first ones to automate and deploy externally. They should be simpler and more beneficial to automate and deliver appreciable cost savings right away. A multi-cloud implementation is a significant change, so trying to do it all at once will result in disaster. Every organization is different and there is no single template to get you there. So set your vision and take your first steps towards it. As you progress down the path, build on the previous steps, course correcting as you go and prioritizing the next steps as they become clear (and they will become clear, trust me).
The most important factor is to maintain momentum by solving your most immediate problems, then use your management system to identify and prioritize your next steps.
Is automation a reality with multi-cloud cloud management?
JL: In private cloud infrastructures, automation has gotten a bad rap for requiring significant resources, both in terms of cost and in the effort needed to customize, implement and maintain. This has led to larger organizations being wary of these types of initiatives and mid-size companies avoiding them altogether.
But it’s important to note that not all cloud automation systems require heavy lifting. There are some that can be implemented in a single afternoon with simple point and click methods. So I’d caution organizations not to dismiss automation in hybrid cloud environments because it’s very much a reality. You need to look for platforms that enable the organization to be agile and deliver rapid, low effort improvements.
Having an integrated automation platform is a key first step on the path to the hybrid cloud. It will enable you to see all the dimensions of the infrastructure from a single pane of glass, automate routine tasks, implement and standardize repeatable processes, identify roadblocks and inconsistencies, understand growth patterns and their impact on your business, and most importantly – prioritize your next step, and the step after that. As I said earlier, once you’re on the path, the next steps will become clearer.
Any parting words of advice for IT organizations as they embark on the path to multi-cloud management?
JL: The notion of a hybrid cloud is no longer a vision, but a reality for many organizations. Getting business buy-in will be essential for a project of this magnitude. In a multi-cloud world, IT needs to morph from managing physical assets and being an organization to be worked around into a responsive, strategic advisor that operates as a flexible, agile service organization to its constituents. With the increasing complexity that comes from applications with many more separately manageable components and physical resources allocated on demand and spread across locations, automation is the only way to keep things sane, manageable and responsive and enable IT success.