The Public Cloud can Teach Your Data Center New Tricks

There is momentum building with public cloud vendors tackling the last frontier, your data center.

By Ravi Lachhman
September 24, 2019

One of my favorite shows on TV today is TLC’s 90 Day Fiance the Other Way. The premise of the show is that TLC follows a few sets of individuals breaking norms in their quest for geographically unconventional love. 

In the technology world, breaking norms or disruption, are pillars of the fast-moving cloud-native world that we live in today.  Becoming the expectation today, as soon as a technology becomes popular that a public cloud provider will provide a commoditized and operationally available version of said popular technology. This is one key potential advantage of moving workloads to the public cloud as we need not worry about scaling or available infrastructure for our new workloads. 

Geographic convention would dictate that the public cloud is most likely not in your own data center; that is the point of the public cloud right? Like 90 Day Fiance The Other Way, public cloud vendors have seemed to have done a u-turn and headed straight back towards your data center. 

Your data center, the last frontier?

In the technology world, innovation and disruption is the norm these days. Very few moments made my jaw drop in the last year [other than Harness] and one of those jaw-dropping moments was at AWS re:Invent 2018. VMWare and Amazon Web Services were on stage together like the viral BMW and Mercedes commercial; once sworn enemies together on stage complimenting the hard work of each other. This was the introduction to Amazon Outposts to the public which introduced AWS cloud services inside you four walls. 

Public cloud vendors have spent years moving workloads outside of traditional data centers. The shift of capital expense [CapEx] to operational expense [OpEx] has been forging forward with companies starting to or completely not owning their physical hardware. Though there must be a good amount of workloads still inside traditional data centers as the public cloud vendors race to the last frontier form them which is your workloads that you yourself self host. 

All in the name of the hybrid cloud, running cloud services inside your environment [or even a disconnected enviroment] seems counter initiative as the convention would be bursting out of your data center, not into your data center.  The “usual suspects” in the public cloud are starting to offer services in your data center.

Who are the players

The three major public cloud vendors [Amazon, Microsoft, and Google] all have flavors of their cloud platforms inside your four walls.  Azure Stack was first on the block with being in technical preview as early as 2016.  Amazon Outposts and Google Anthos round off the top three public cloud vendors. Honorable mentions from other cloud providers such as Oracle Cloud at Customer and IBM’s Hybrid Cloud are also compelling solutions. 

The AWS Solution can be a fully managed solution including hardware, software, and services in your data center. The Microsoft Solution can be installed on a series of certified partner hardware.  These solutions are recent additions to cloud vendor’s lineups and would not be offered if not relevant today.

Why is this relevant?

Minus regulatory requirements or technical requirements such as edge/disconnected environments, cloud cost containment is a big driver. Optimizing workloads for cloud cost containment can be difficult.  Having a more predictable model using your infrastructure can be enticing for enterprises. For more risk-averse organizations or organizations that are slower to adopt public cloud offers, injecting new life into the on-premise stack can be one way towards a hybrid cloud.

Let’s say you have an existing investment in hardware that can be re-purposed. All of these “new-fangled” workloads are headed towards Kubernetes. As we completed our six-part blog series on Kubernetes, running Kubernetes is challenging. As a product and service, Google Kubernetes Engine [GKE] helps solve operational complexity such as identity management and scaling on your Kubernetes cluster(s).  What if you had the ability to run GKE’s management plane in your data center? Well, look no further than GKE On-prem as part of the Anthos platform. The ability to run a cloud provider’s layer on your infrastructure can be a viable option.  

In terms of cloud migrations, the Five/Six R’s from Gartner are driving principals when strategizing workloads to move to the cloud.  With running cloud vendor software on your infrastructure, the first two “R’s” which are rehosting and re-platforming become easier with infrastructure you are already familiar with and infrastructure that meets your organization’s standards. 

Harness, your omnipresent friend

As you are going through your cloud migration journey, Harness is there to make sure your pipelines are robust enough to handle heterogeneous workloads and infrastructure. A shining example would be our very own migration story which we migrated cloud providers completely in less than two hours. 

The Harness Platform is pretty plugged into multiple cloud providers and on-premise destinations for your Continuous Delivery pipeline. Even if perceived to be going the opposite way with one of the newer on-premise public cloud solutions, Harness Pipelines are flexible enough to go either direction.  At Harness, we are omnipresent where your infrastructure is. 

Cheers,

-Ravi

➞ Back to Blog

Leave a Reply

avatar
  Subscribe  
Notify of