Trends in Software Defined Data Centers


I recently met with a regional cloud service provider (CSP) that has adopted provisioning on demand as their IT model. They spin up applications on demand, having virtualized most of their infrastructure and have developed tools to automate the provisioning of applications and servers for customers/tenants through a self-service portal.  Rather than build-out and manage more and more physical infrastructure, with associated time and expense, the CSP is adopting the concepts of software-defined data center (SDDC) that builds on virtualization (of software, networking and storage) to offer software and network services for many different clients. More importantly, the CSP is also catering to needs of DevOps and IT architects – both internally, and externally for its tenants/clients by enabling true self service through automation.

It is important to note that SDDC is a way to deliver data center services efficiently, reducing errors and thus influencing the costs associated with operating an infrastructure.

The benefit for tenants is that they do not have to build the infrastructure themselves and can consume the computing, networking, and storage resources, simply by renting and paying for them as they consume them via cloud – private, public or hybrid. Tenants also benefit from the best practices of self-service, quicker provisioning, reduction in costs associated with errors in manual configurations. The advantage for CSP is that in addition to quicker provisioning for its clients, and the ability to scale software, network and storage, they differentiate themselves from competition and generate additional revenue through value-added management and automation tools. Of course, by allocating the operational cost of this infrastructure over many tenants to bring economies of scale allows the CSP to lower the operational costs for its tenants.

In addition to virtualizing the data center, SDDC also changes the way IT architects design and think. The goal is to adapt to demands of continuous delivery of applications, reducing the costs of deployment while adopting best practices. This leads to incorporating and adopting microservices, automation, templates and DevOps practices.

[You might also like: Cloud Load Balancing – Does your provider have what it takes?]

As we saw with service orientation earlier, with the next iteration, the software components are evolving into microservices with containers as a delivery mechanism. Microservices are becoming more prevalent due to increased migration to cloud to enable continuous delivery to update applications and software. This capability gives organizations an edge in time-to-market, keeping up with both their business needs and the rate of change in technology.

DevOps as a concept and a process for product delivery allow organizations to reduce the time it takes to incorporate feedback into product development.

In an environment where tenants may share the resources, a wrong configuration change of a single tenant may impact all other tenants – severely impacting an application’s SLA and availability. In addition to architecturally separating these tenant instances, it is also important in terms of administration to have the necessary tools to automate error-prone tasks.

This is important in terms of provisioning on-demand resources but even more relevant in terms of cost saving while reducing error prone tasks by automating the process. While tools like Chef, Puppet, and Ansible greatly help automation, we can expect new products to provide programmability through APIs, out-of-the-box support for automation and templates to achieve self-service in an SDDC.

In addition to a virtualized infrastructure, SDDC introduces automation principles, programmability, policy-driven provisioning of application environments through out-of-the-box templates and enables application owners to define and manage their own environments, based on performance, security and business continuity requirements.

Learn more about vDirect, Radware’s ADC Management Orchestration Plug-In here.

Learn more about Radware’s Operator Toolbox here.


Read “Keep It Simple; Make It Scalable: 6 Characteristics of the Futureproof Load Balancer” to learn more.

Download Now

Previous articleProtecting Cloud-Based Applications – Not What You Thought
Next articleYou Cannot Do Application Delivery Without Security
Prakash Sinha is a technology executive and evangelist for Radware and brings over 29 years of experience in strategy, product management, product marketing and engineering. Prakash has been a part of executive teams of four software and network infrastructure startups, all of which were acquired. Before Radware, Prakash led product management for Citrix NetScaler and was instrumental in introducing multi-tenant and virtualized NetScaler product lines to market. Prior to Citrix, Prakash held leadership positions in architecture, engineering, and product management at leading technology companies such as Cisco, Informatica, and Tandem Computers. Prakash holds a Bachelor in Electrical Engineering from BIT, Mesra and an MBA from Haas School of Business at UC Berkeley.


Please enter your comment!
Please enter your name here