We're living in the age of Everything-as-a-Service.
But there are still a few things in the tech world that don't work well when delivered as a service.
From software (SaaS) to application platforms (PaaS) to computing infrastructure (IaaS) to data storage (DaaS), almost everything you require to meet your IT needs can now be obtained as a service – and in many cases, obtaining it as a service is more cost-efficient than doing things the old, on-premise way.
But it's important to recognize the exceptions to the Everything-as-a-Service, or XaaS, paradigm.
Here are some examples of IT resources that still can't be delivered and consumed very effectively as a service.
You can outsource most of your infrastructure to the cloud using IaaS.
But at the end of the day, your employees need a screen, a mouse and a keyboard in order to access remotely hosted resources.
OK, maybe you could get away with a tablet or phone as the portals your employees use to connect, but I still don't think a touchscreen comes close to matching the productivity of an old-fashioned keyboard when it comes to doing real work.
And anyway, if you rely on mobile devices for your end users, that's still a type of on-premise infrastructure.
If you use software that requires a lot of graphical computation, you'll be hard-pressed to run that software in the cloud.
The reason is that it's difficult for public Internet connections to handle the data throughput that you'd require to run graphic-intensive applications on a remote server and deliver the display in real time to your workstation.
While you can run some graphic-intensive applications on servers on your local network and access the display from a workstation, you still can't run things like sophisticated games or video-editing software as service in the public cloud.
You need local GPUs to make them work.
You can consume some network resources as service.
Network-as-a-Service, or NaaS, providers use network virtualization to deliver sophisticated networking configurations to clients without requiring the clients to set up complex networking hardware and software themselves.
But those clients still have to have some kind of network infrastructure in place.
You can't completely turn the network into a service.
If you run any kind of hardware on premise, you need an operating system to power that hardware.
There's no good way to consume operating systems as a service.
Sure, technologies like PXEboot have long made it possible to boot to operating systems hosted somewhere else on the local network.
But this approach doesn't work well if you want to host your operating system in the cloud and have it delivered as a service.
Some companies now use the term "operating system as a service."
However, they're usually referring to management services and automatic updates for operating systems, not a service that delivers the entire operating system from the public cloud.
XaaS is a great model and a handy way to gain access to compute and storage resources without having to set up and maintain them yourself on-premise.
But there are still some things that just don't work under the XaaS paradigm.
Recognizing the exceptions to the XaaS trend is important for MSPs because the exceptions are the types of services that only traditional managed service providers can truly deliver.
Items like workstations and network infrastructure require more hand-holding and configuration to set up and maintain than customers can acquire from a generic XaaS provider who only wants to deliver software from public cloud servers.