Safely! I would imagine is the resounding answer.
Inherent to the nature of cloud technologies, are that advances are borne from experimentation - tackling existing and emerging challenges using non-traditional approaches. So it seems clear to me, that anyone looking to take advantage of the accelerating improvements in these areas needs to adopt an approach of cloud-trialling and cloud-experimentation.
How close you move towards the edge of your comfort zone will be dictated by the constraints of your organisation (real, perceived or inherited).
Don’t get me wrong, I am not advocating reckless abandon. However, it is readily apparent that there is a continuum of approaches emerging for jumping onto the cloud train, ranging from the tactical and impulsive to the strategic and considered. And even with government mandated to use the AoG utility cloud, the directives are softening. Cloud transformation is all about leveraging new and emerging models for business differentiation and advantage. So the spot(s) on the continuum on which you board has a large bearing on the potential benefits that can be derived.
By way of examples, I’d like to discuss a few different approaches to jumping onto the train; AWS Management Portal for vCenter, Hotlink, Ravello and VMware vCloud Automation Center.
AWS Management Portal for vCenter
Recent announcements around the release from Amazon Web Services of a plugin for VMware vCenter have elicited dichotomous responses that vary from ‘wolf in sheep’s clothing’ to ‘enlightened consumption enabler’. The plug-in requires some trust relationships to be established between vCenter and your AWS account and from that point you can deploy workloads directly into AWS. For me, this is a fantastic tactical enabler that encourages experimentation and develops familiarity with AWS. Fair enough though, there are some warnings about how one might migrate workloads back on-premises (if you require the plug-in to get in you are probably not comfortable with a CLI interface to migrate in the other direction). There are also challenges around consistent networking configuration across multi-VM deployments, software licensing, backup and more. However, I think that the silver lining to those objections is that at least you will now be forced to think (albeit tactically) around how to address the challenges as they come up (aka attain knowledge and experience). So I would say by all means leverage this approach, just go in with your eyes open and holding the hand of a friend.
Hotlink - checkout time available at the Hotel California
A wonderful aspect to the evolution of the cloud market is the new ecosystem of vendors that emerge to bridge the gap between what is available and how customer would like to consume those services. Ravello and Hotlink both fall into this category. Hotlink is a more traditional approach to migrating service into the cloud (AWS in this instance). Hotlink Hybrid Express is really a corporate ready version of the AWS vCenter Portal plug-in. It allows you to migrate workloads into and out of AWS from within vCenter, and the wizards guide you through additional aspects like the mapping of networks and security groups. They also have another tool Hotlink DR Express that lets you store VM snapshots in AWS on a regular schedule. The snapshots can then be used to run DR instances inside AWS in the event of a failure.
Ravello - encapsulate to abstract complexity
Ravello Systems take a more non-traditional approach that uses encapsulation techniques. Essentially, a multi-VM deployment, along with its network connectivity dependencies is encapsulated using a combination of binary translation (back to the future!), software defined networking and storage overlays. What is more, the set of VMs that you run in the cloud are consolidated onto a fewer number of larger cloud provider instances, which results in efficiency and performance. This is definitely exciting from a technologist’s perspective, however it also addresses the challenges of retaining the fidelity of an environment as it moves between clouds, avoiding time consuming and manual reconfigurations. The cost model is also interesting as it overlays their functions onto cloud based pay-as-you go models so that you pay a per hour price for the Ravello system plus the cloud components on which it runs.
vCAC - a safe pair of hands
Then we come back to what the mother-ship is able to provide. VMware vCloud Automation Center is undoubtedly the future for most organisations of any significant scale. It enables reaching out into other clouds but it does a whole lot more too, including orchestrating the workflows that make a VM container useful (e.g. OS, patching, applications, updates, networks), adding lifecycle management to improve overall utilisation of cluster resources and presenting a catalog of services (which if done right, represent the services that the business wants to consume). In our experience, leveraging this approach is the most considered and we have seen that it can be a slower process, since there are considerations of changes to licensing models, evaluation, design and commissioning. So whilst this is most often the right near term destination (on the cloud ‘journey’), there is good reason to consider immediate term approaches like the ones discussed previously.
Finding the right approach
At ViFX, our reason for being is to be agents of cloud change, inside and outside the boundaries of your organisation. And we look to do this for organisations that are seeking to leverage existing, new and emerging approaches for their business advantage. We want to be part of that change with you and for you.
For these reasons we have continued to develop the models, frameworks and approaches that stood us well through the years of virtualisation, virtual infrastructure, private cloud and now into the hybrid cloud era. This approach can be summarised as considering challenges and opportunities in their wider context of aligning to business goals or simply the context of the IT systems with which they interact. Also we remain open-minded and agnostic so that the merits of different approaches and solutions can be fairly evaluated in the context of the requirements, in order that a robustly defensible recommended solution and adoption roadmap emerges.
Please bear in mind this is just a sampler of myriad approaches and technologies that exist. There are many more that should be considered (and we’re looking at them). But the point is not to exhaustively inventory all of them (and then start again once you are finished to catch all of those that emerged since you began). The point rather is to illustrate that there is an approach out there that is right for every organisation and depending upon the challenge immediately in front of you (e.g. urgent development needs, reactive capacity, individual projects vs business service SLAs, internal cost management). You just need to map your needs, constraints and desires to the right approaches available in the ecosystem. And sometimes you just need to roll up your sleeves and get stuck in.