Join David Linthicum for an in-depth discussion in this video Building an ops process, part of Planning a Multicloud Solution.
- [Presenter] So let's talk about operations in multicloud. So what's important here is that multicloud operations, or ops, is about planning and leveraging the right tools that support automation. To be successful in doing this we need to automate virtually everything. We're trying to remove the humans out of the process of operating our Multicloud because it's too complex, too many things are occurring, it's event driven, we have lots of things that are happening all the time than need to be addressed, self healing processes, we need to be able to reboot servers automatically.
We need to be able to auto-provision storage, auto-provision compute, that you're really trying to automate virtually everything about operations. So you need to do this before deployment of existing or net new multicloud operations. So planning needs to occur, and so if you do not plan, you're going to plan to fail. I know that's an old term, but in the world of multicloud operations, it's absolutely true, because if we don't think about assembling the right sets of technologies in automation and orchestration, the operator systems, then ultimately we're going to run into walls.
We're going to hit a tipping point very quickly, where the complexity is overwhelming for the humans that are operating the system, and things are going to fall down. So, ops, operations, should be continuously improved, and indeed can be part of DevOps, so you have to remember that right now we're linking development and operations together. And so developers are focused on operations, and operations is focused on development. So the wall that existed between development and operations are falling down.
And the ability to do infrastructure as code, the ability to test or build reliability into the applications, build security into the applications, governance into the applications, the ability to test in the right ways, so we're focused on providing the right operational services as well as the right application and development is really where it's at right now. So you can't separate the two. We're starting to focus on operations but operations is linked with the development and development is linked with operations. So, we have security operations, have to consider, and this is about securing systems, so we're leveraging the right tools and technology to protect the system as it's in production.
Gov Ops, the ability to provide governance, and we know about governance from the previous video. Performance Ops, which is the ability to enhance performance or monitor performance and tune for performance in an ongoing continuously improving way, so we're not just monitoring performance when the performance start to degrade, we're launching additional systems, or picking more powerful platforms, or CPUs directly from our public cloud that we're leveraging, but we're able to do this in an automatic way, so if it detects that performance is moving off of the SLAs, then ultimately, the performance of the system will be tweaked and auto tuned and auto scaled directly from performance operations.
So we're doing monitoring and management, we're looking at performance, and we're constantly improving our ability to provide performance. Application operations, the ability to understand how applications fit into operations, so when they're promoted into a staging area, the types of tests that occur in terms of making sure that the applications are ready for production. And then ultimately within operations, the ability to monitor the applications as they're running, just like we monitor the databases and monitor the storage systems, and monitor the CPUs. Now this is rather new, because typically we monitor at the platform level, which supports the applications.
And now we're monitoring the applications, as well as the platforms, as well as the storage systems, and we're doing so because the applications themselves, if they're not running correctly, they're not providing the users with the services that they need, and they need to be corrected. So, monitoring and management tools that you pick out there today, need not only to monitor the platforms, storage and compute, and networking, but also the ability to monitor the application levels as well. Data Ops, the ability to monitor the databases and storage systems out there to ensure that they're not filling up, to ensure that they have enough resources to run, the cache sizes are correct, they're tuned correctly, all these things basically have to be part of your operational infrastructure.
Your monitoring and management and your ability to automatically correct systems that are occurring with the data. And then other ops, you may have IOT based systems. You may have big database systems. You may have odd things such as factory floor monitoring that may not be part of all different kinds of domains. And those have to be included in your operations as well. Obviously if your multicloud is communicating with robots out on the factory floor, then you're going to have that as part of your operations, part of your monitoring, part of your management, part of your restarting, your self healing, all these things that may need to occur to keep those things up and running.
So again, it's not only what's important here in terms of common patterns, you're typically going to have Sec Ops, Gov Ops, Perf Ops, App Ops, Data Ops, but what's specific to your particular problem domain. And there's going to be lots of odd things that are going to be part of the deal, and you need to include them in your operational things as well. So automation, tooling, and people are where you need to focus. Automation being first and foremost. I'm trying to automate everything here so that human beings don't have to get involved in this. And therefore it's going to have a greater chance of success.
And to automate you need to pick the right tools. The ability to leverage the right technology. And then people have to get involved from time to time, and they have to be trained and have the right skill sets so they can do security and governance and performance and applications and data. All those things need to be there. And you need to have particular skill sets so they know how to get in there and fix the database, they know how to get in there and fix the governance system, however, those should be at instances in time, typically, when human beings get involved, at odd times when things go wrong, very wrong, but far as operational systems and basically stuff that runs on a daily basis, we need to focus on automation, need to focus on tooling.
Instructor David Linthicum covers building out operations and development processes; planning governance of resources, cost, and security; putting together a strategy; and staging deployment and testing. He also examines several use cases featuring multiple service providers, showing how these solutions are architected in the real world.
- What is multicloud?
- Use case studies
- Common patterns and problems
- Selecting cloud technology
- Building an operations process
- Building a development process
- Multicloud strategy and planning
- Deployment and testing