IT infrastructure powers the applications, seen and unseen, that surround us so much we’ve become acquainted to them. A great feature of our brains (is it?) is that it tends to do thermal / damage control – when there is too much complexity going on, or when our thoughts sometimes get too intricate, it will magically divert our mind to something simpler, so that we don’t overheat or melt down.
But whether our brains accept or refuse to see it, IT infrastructures are getting more and more complex. We could refuse that thought and recall the good old days of server-client architectures, a time when things were simpler, but realistically, the complexity has been around for at least two decades, if not more. And the technology layering doesn’t helps either: we are talking about micro-services, workload mobility and machine learning, but someone’s still has to feed and support the mainframe living in the basement of the data center.
Challenges of Managing Infrastructure at Scale
If we put aside “monsters” living in their own lair, IT infrastructure is ubiquitous, be it servers, networking or storage, and we’re not getting into details! To this three pillars we must add security, data protection, data management and much more. We have no statistics, but organizations and infrastructure engineers are bound to use multiple management interfaces to get work done during their day. Advanced / specialized team members can rely to scripts and code to get around the rigidity of manual work and get work done efficiently.
But even this becomes a double-edged sword: scripts / code and the tools used to execute them need to be documented, referenced and maintained centrally, else it becomes lore only known to a few, and when those privileged few fade away, so does the lore. If it wasn’t enough, APIs used to make those scripts run can change, breaking cranks that used to work perfectly well. And while this may work in smaller environments, doing this at scale becomes a nightmare. As more components are added, and as the technology / platform scope increases, maintaining those scripts because a job of its own. Are scripts bad? No, they’re not – it’s not a problem of technology but management and control.
Why We Need Agnostic Infrastructure Management
Managing your entire infrastructure from a true single pane of glass seems to be the Holy Grail of infrastructure management, seemingly attained so many times, but never in true reach. There was always something missing: when we had a comprehensive view of the metrics, we didn’t have a way to take action. When we could orchestrate things, we wouldn’t get monitoring. And when we had the cost view, we would miss either monitoring or the orchestration engine. While this may oversimplify things a bit, it is a view of where we’ve been for a long time.
But do not cloud operators offer all of these capabilities today? Sure, they do – although I would dare to say quite badly – and most enterprises are operating hybrid infrastructure anyways. They need a single control plane that allows infrastructure to be managed seamlessly, whether it is the on-premises VMware footprint, or the cloud workloads scattered across the major public cloud providers. They need a management plane that abstracts underlying technology and makes infrastructure a seamless continuous plane, regardless of changes happening under the hood, while putting applications at the heart of operations. Finally, they also need to orchestrate all the interactions between infrastructure components, data movements, and provisioning / de-provisioning workflows, ideally without having to maintain a complex code base.
Morpheus Data – Agnostic Infrastructure Management in Practice
At Cloud Field Day 11 we had the opportunity to listen to Morpheus Data, a company focused on building an infrastructure agnostic application management and orchestration solution.
Morpheus have approached those agnostic infrastructure management prerequisites in an interesting way. It provides various capabilities such as IT self-service, access management / cost control, automation, cost optimization / financial visibility, and hybrid cloud management. It takes organizations from the foundational stone of IT infrastructure into lean, true cloud operations with automated provisioning, and also adds the financial perspective of cloud spend – not just in terms of provisioning costs, but also looking at actual costs, and giving insights about spend and potential data migration opportunities to competing clouds.
Morpheus Data just doesn’t looks at infrastructure from an IT operations perspective. The company acknowledged that different personas in the enterprise need access to infrastructure management, but each in a different way: developers or business users have different needs than infrastructure managers, therefore the platform allows each of those personas to get a customized experience. For example, business users would get access to a self-service catalog, just like they would in the public cloud.
One of the big challenges we highlighted earlier is the lack of integration between infrastructure components and the need to develop scripts to bridge the gap between systems, with more or less luck and outcomes. Morpheus Data also clearly understood that if they wanted their platform to be successful, they had to own the technology integration stack. This is by no means an easy task, but Morpheus Data includes almost 100 codeless technology integrations. This means no code, just workflow creation directly within Morpheus Data.
Advanced users or organizations using bleeding-edge products should however be able to palliate the lack of a codeless integration by running their own custom code in some of the workflows.
Infrastructure management can very rapidly become a nightmare. We (IT professionals and organizations) sometimes forget to take a step back. We’re used to incremental scope creep, where one tool gets added on top of a new one, then three months later a new vendor is added, then we inherit something previously managed by someone else. We found ourselves buried with work that makes no sense, and while some try to automate as best as they can, we sometimes end up missing the bigger picture.
Infrastructure must be managed, business outcomes must be delivered, in other words: “the spice must flow”. But it should flow in an efficient manner, and we should really refocus on the outcomes and efficient management, rather than try sticking things together with duct tape and scripts.
Morpheus Data was one of the highlights of Cloud Field Day 11. The solution does so much that it’s almost overwhelming, and yet organizations can start with simpler functions and gradually ramp-up as they get more comfortable. And they can very probably also consult with Morpheus Data to establish a comfortable bridgehead.
The part we enjoyed the most though was related to cost management. This is such a big deal for organizations, because it can actually show them what the actual spend looks like, what provisioning some workloads on the cloud (or storing some data) will cost, and what it means if we move it somewhere else. True, some organizations have special rates or deals with cloud vendors. Nevertheless, many in the organization (except finance / procurement) lack the clarity about financial cost, and being informed upfront will hopefully help organizations make more informed choices.
The following videos were recorded at Cloud Field Day 11:
2. Technical Introduction to Morpheus Data Hybrid Cloud Management
3. A Unified Approach to Clouds, Containers, and PaaS Services with Morpheus Data
4. Cloudy with a Chance of FinOps: Hybrid Cloud Cost Management with Morpheus Data