At the end of each of our webinars, we hold a Q&A session where the audience can ask questions that arose from the webinar content. This blog series highlights some of those questions to expose them to a wider audience.
This week the questions come from the webinar, Beyond Cloud Automation - Reshaping Your Cloud Strategy, where Forrester Principal Analyst Dave Bartoletti and Embotics CTO Scott Davis discussed key trends in Hybrid cloud computing, and how they could impact your business strategy.
Q1: What is the future for the private cloud?
Dave: I think the future of the private cloud is as a development platform. We have seen a lot of initial private cloud efforts over the past couple of years tht are either stalled or in the worst case, failing and that’s because organizations have been challenged in determining what their private cloud is supposed to be. Was it a competitor to the public cloud? Or was it a hedge against additional public cloud use in the enterprise? Also, a lot of those projects have stalled only because of the innovation pace in the public cloud.
There’s no problem calling what you’re doing your "data center," or "a private cloud" if that helps justify spending for it and maybe if it helps a little bit with IT efficiency, but if it's advertised internally as an alternative to the public cloud, developers that actually have experience in the public cloud are going to see the difference in scalability, breadth of services, and in the consumption models.
That said, what we're seeing now is the second wave of successful private clouds that are not just infrastructure platforms, they’re development platforms.They’re not just open-stack, they are open-stack plus Cloud Foundry, or plus development tools on top of it, going back to our roots in the private cloud.
If it’s to help our business and development teams create amazing software faster, then private clouds that focus more on that experience, the services and tools that developers need to build more effectively, and less on, ''We're doing private cloud just to drive up data center efficiency," they are being seen as more successful.
I also think that the future for private cloud is going to be impacted by the evolution of API consistent public cloud platforms available on-prem, so things like Azure stack, and Bluemix local and on-prem versions of the leading platform-as-a-service platforms. I think private cloud is growing up.
Q2: What are your thoughts on the future of OpenStack as a public or private cloud platform.
Dave: I think OpenStack by itself is not enough to create a private cloud right, it has been difficult and time consuming to date but it’s getting much much easier. So I think two things are going to happen. First, we are going to see an increased demand for remote managed OpenStack deployments and appliance based OpenStack private clouds where enterprises really won’t have to take the initiative of building it themselves, and maybe not be operating it downstream. And second, I think we’ll see much tighter integration between the OpenStack stack, and container services, and development platform services.
It’s not just unique to OpenStack the PaaS market is also faced with challenges. What is Cloud Foundry today? What is OpenShift? Some people see them as container orchestration platforms as much as they see them as end-end middleware platforms, or integration platforms so all of these markets are being disrupted by the advent of both containers and serverless computing and that’s going to apply to OpenStack as well.
Scott: I do agree with Dave that the private cloud is, a development environment, but I also see private cloud as being used for functions that really provide competitive differentiation for a particular organization. I am seeing all of the back office business operations functions moving faster to the public cloud, and not being developed and managed individually in organizations.
Dave: I think it’s part of a broader change in what open source means to enterprises. Open source historically going back to early linux days, was a low-cost reaction to expensive proprietary software. Open source today is not a reaction to expensive preparatory software. It’s speed. People want it because they can grab these components and do something quickly with them.
I think in the early days OpenStack was pitched as a hedge against big closed proprietary Amazon behemoth's. People don’t necessarily feel frightened or locked into a platform that keeps getting bigger, better and cheaper all the time. So as a reaction to public cloud platforms, I don’t think it wins. As a place for enterprises to build these great new database and application experiences like Scott said, were for whatever reason it makes more sense for them to do it in-house. That could be that our CISO was just not ready to certify by public cloud, we just built a data centre, We’ve got tons of equipment that we want to reapply to something that really differentiates our business, that’s where the success is going to lie, not just in an incremental add-on to virtualization.
Q3: Most of the cloud platforms have their own management tools. What are the advantages of going with a third party management platform?
Scott: The fundamental advantage is a lack of vendor lock-in. When I talk to customers today most of them have on-prem footprints, usually VMware, they have applications running in-house, and they want to use more than one cloud, and don’t want to be locked in to a single vendor.
Each of the vendors tend to have tools that go deep and manage their particular offering well and do very little for other offerings, they're not really heterogeneous. So a key advantage of a third party management platform is, if you pick the right one such as Embotics, it goes deep on the platforms that it supports, and you can now handle this multi-cloud world far more effectively. You can create images and manage with a consistent point-of-view, and manage the costs between them.
Q4: What is the concept of serverless computing, what are it's pros and cons compared to microservices, and do you need both a private cloud platform and a microservices platform?
Scott: The way to think about serverless computing is as code-as-a-service. It's a higher level abstraction than you've had before. Microservices and containers are about every individual function of an application. Each dedicated function is executed in a dedicated container. It does one function, one function well. You are still, in creating the application, managing and scaling each one of those containers.
With severless computing, the application infrastructure is actually becoming a service itself. You're uploading code, which is a function, to your cloud provider, and when an API call is made, instead of that API call being funneled to an existing microservice at some IP address it's triggering a new instance of execution at the back-end that's going to load and run your code. The scaling of the application infrastructure is offloaded as a service to the cloud provider.
Dave: We're in the early state of this so the naming is terrible. It is serverless from a developer's standpoint and this is how we're marketing this to development teams, or how the industry is marketing it to development teams. Really the idea is the ultimate level of abstraction for developers, where now I don't even have to put my functions in a container and even think about how they're going to run packaged. I just send the code to a service and it executes, let alone creating a virtual machine. Of course all of this runs on a server, either in a VM or in a container or on bare metal. That's not the point from the developer's standpoint, so Forrester calls it "functional pipeline programming" because it's really much more akin to database triggers than it is to any traditional way of deploying infrastructure.
Q5: What are you hearing from clients who have successfully automated, and how they breaking down barriers to automation?
Dave: You need to start small but for each one of these small automations, you need to tie it to a business outcome. You need to articulate the value of that automation immediately in something that matters to the business. We in IT get stuck a lot of times articulating the value of automation in IT terms. "We have faster provisioning time", so what? How is faster provisioning time helped the business or a development team release software faster or open a new office or acquire ten more customers? The more that we can articulate the value of these small incremental automations the better.
The important thing to do is, chop your automations into smaller and smaller bits as you're chopping your code, and your DevOps process into smaller and smaller bits. And then for each one, advertise. Advertize within the organization the business benefit achieved by automation. That's how you'll fund the next phase.
You can watch the complete webinar with Dave and Scott by clicking the link below.