On-Prem to the Cloud to Edge – Why Do I Have Servers Again?
Is the edge important to me and my business? Chances are, yes, the edge can be very beneficial to your business. Let’s walk through how we’ve evolved from on-prem to cloud to edge, and what this could mean for your business.
Aren’t We Done with Servers?
For years, we had racks and racks of servers in data centers and they ran all our infrastructure and applications. It may not have been the best approach to every business case, as they required large capital outlays at regular intervals, cooling and power infrastructures and reliable connectivity, but it worked pretty well for most of us.
The emergence of the cloud has helped us to migrate away from those physical servers and their associated commitments; we changed our cost model by taking advantage of the granularity of resources and began paying for resources based on actual consumption. Servers and racks were retired, and businesses were left to focus their on-prem infrastructure on the last piece of the puzzle – connectivity. Our business leaders loved us – shifting from large cash outlays, based on estimated usage models, to a consumption-based model that directly reflected resources that were used within the last billing period.
Fast forward about a decade. Application modernization has shown us how to decouple workloads to utilize containers, serverless computing and other platform-based (PaaS) products. The racks of servers we previously maintained have been replaced with much smaller deployments, if we even have any servers at all.
Step to the Edge
The next stop along on our journey had brought on the revolution of edge computing; the philosophy whereby we connect people and endpoints to the resources they are attempting to access by physically limiting the distance between them. What started with Content Distribution Networks (CDNs) that allowed us to create replicas of our content at endpoints around the globe, has now evolved into extending the barrier of the public cloud onto our physical networks.
Wait. Did he just say that the public cloud is extending ONTO our networks?
Yes. Our physical networks are the next target of the cloud, with the goal of providing the best user experience possible. Share on XExtending the boundaries of public cloud infrastructures onto our physical networks seems at odds with everything we have done to embrace cloud adoption. The cloud promises us less maintenance, less overhead and rapid deployment times. So, why are we bringing servers out of the cloud and back onto our networks?
It helps to look at this through the context of your applications. Let’s assume you have dozens of network-connected video cameras, and you want to use those cameras to analyze streaming video data in real time, and apply a machine learning (ML) algorithm to that streaming data to identify the resources that you have specified. In a traditional cloud model, that video data would stream in its raw format across your network, across your connection to your cloud provider, processed by those cloud resources, and then the relevant data must stream to its endpoint. All that travel time adds latency and decreases the responsiveness of an application to a user.
What if data transmission and ML processing happened on your local network, and was then asynchronously synchronized to your cloud-based data stores? This is a great example of how the edge can be extended onto your network to improve application response. Local data storage provides rapid access to information, ML processing happens closer to your users, and as such, your application response times are decreased by consuming resources from an appliance located on your network.
What this Means for Business
Is the edge important to me and my business? Chances are, yes, the edge can be very beneficial to your business. Architecting solutions that take advantage of the cloud and the edge, and your local network, using connected appliances that are managed through cloud management fabrics bring application response closer to your control, without requiring a new operational plane to manage.
If it helps, look at it this way – these are not really servers in the way we traditionally think about servers. These devices are network appliances that happen to be on our physical network. Yes, we are paying for them, and maybe that payment model is different than we have grown used to for cloud resources, but these appliances have almost no impact on our operational maintenance models. Bringing the edge to our networks can be plug and play – if you architect it correctly.
