cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post

Breadth is the New Depth: The Value in Knowing a Little Bit about the Entire Cloud Stack

Level 10

Cloud fixes everything. Well, no it doesn’t. But cloud technology is finally coming out of the trough of disillusionment and entering the plateau of productivity. That means as people take cloud technologies more seriously and look at practical hybrid-cloud solutions for their businesses, engineers of all stripes are going to need to expand their skills outside their beloved silos. 

Rather than focusing only on storage or networking or application development, there is great value in IT professionals designing and building cloud solutions knowing a little bit about the entire cloud stack.

The National Institute of Standards and Technology (NIST) defines cloud computing as “a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.”

We call it a “cloud stack” because of all the components built one on top of another. This includes elements such as networking, storage, virtualization, compute, load-balancing, application development tools, and more specific to operations, things like user account management, logging, and authentication services. These are all built right in to the IaaS cloud.

But when looking at the overall picture, the overall cloud stack, IaaS exists as the foundation for the Platform as a Service, such as development tools, web servers and database servers, which in turn serves as a platform for Software as a Service, such as email and virtual desktops.

So when an IT professional is looking at a cloud solution for their organization, regardless of their background and specific area of expertise, there’s a clear need to be able to understand a little bit about networking, a little bit about storage, a little bit about virtualization, even little bit about application development. Sure, there’s still a need for experts in each one of those areas, but when looking at an overall cloud (or more realistically hybrid-cloud) initiative, a technical engineer or architect must understand all those components to some extent to design, spec, build, and maintain the environment.

I really believe this has always been the case, though, at least for good engineers. The really good IT pros have always had some level of understanding of these other areas. Personally, as a network engineer, I’ve had to spin up VMs, provision storage and work with validation platforms to one extent or another from the very beginning of my career, and I don’t consider myself that great of an engineer.

When I put in new data center switching and firewalling solution, I’m sitting down with someone from the storage team, Linux team, Windows team, virtualization team, and maybe even the security team. Often I need to be able to speak to all of those areas because of how, when it comes down to it, our individual sections of the infrastructure really all work together in one environment.

Cloud is no different

All those components still exist in a cloud solution, so when IT pros look at an overall design, there’s discussion about network connectivity, bandwidth and latency, storage capacity, what sort of virtualization platform to run, and what sort of UI to use to deliver the actual application to the end-user. The only difference now is that the cloud stack is one orchestrated organism rather than many more disparate silos to address individually.

For example, how will a particular application that lives in AWS perform over the latency of a company’s new SD-WAN solution?

And in my experience, I see hybrid-cloud approach more than anything else which requires very careful consideration of networking between the organization and the cloud provider and how applications can be delivered in a hybrid environment.

I love this, though, because I love technology, building things, and making things work. So the idea that I have to stretch myself outside of my cozy networking comfort zone is an exciting challenge I’m looking forward to.

Cloud doesn’t fix everything, but organizations are certainly taking advantage of the benefits of moving some of their applications and services to today’s popular cloud provider platforms. This means that IT pros need a breadth of knowledge to provide the depth of technical skill a cloud design requires and today’s organizations demand.

6 Comments

Sigh.  I'm sorry for being the gad fly, the continual "cloud nay-sayer."  But before printing--even in jest--the words "Cloud fixes everything," it should at least be available and reliable.  And it's not.

NetPath shows it.

Image 1:  Path to Microsoft's cloud-based onenote.officeapps.live.com--all resources are currently down, but were available earlier today:

pastedImage_1.png

Image 2:  Path to Microsoft cloud's powerpoint.officeapps.live.com--unavailable at present:

pastedImage_2.png

Image 3:  Microsoft officeapps.live.com--down:

pastedImage_3.png

This goes for cloud-based URL's supporting Microsoft's products:

  • Powerpoint
  • Visio
  • Word-edit
  • Word-view
  • Officeapps
  • Broadcast

I'm open to the idea that maybe routing, firewall rules, NetPath configuration are configured incorrectly.  Except when traffic flows properly and I can confirm the apps are available in the cloud personally.  And then NetPath shows them red, and I can't run the apps.

That suggests the rules and NetPath are built properly.

Nope.  You can't convince me that the cloud, at least these Microsoft cloud-based services, are a good solution.  Maybe one day.

But not today.

Level 20

I'm glad there's some support for AWS now in Orion.

MVP
MVP

Good article

Level 21

rschroeder​ sorry to see you have had such a poor experience 0365, we use 0365 extensively and have not had any outages that immediately come to mind.  When you are seeing those issues in NetPath are users also not able to access those resources?

Level 21

The only difference now is that the cloud stack is one orchestrated organism rather than many more disparate silos to address individually.

We are treating the cloud as though it's different in this regard but I would argue that it isn't.  All of these things have always been all part of one organism.  They may not have been orchestrated the same and we may not have treated them like they were all part of a large whole but that doesn't change the fact that they were.

I think one of the other things that has been changing and not nearly recognized enough are the applications.  Applications now relay on so many disparate resources working together as one and it's important that those resources are all treated as part of that larger whole application and doing this isn't easy.  You really need to map our your application in the same way you used to map out your network; this way when failures occur you have a schematic to work from.

Frequently, yes.  I'd hoped the apparent unavailabilities listed by NetPath were simply errors on my side, or on our firewall configuration.  But that turns out not to be the case.