Jim Grubb, VP Marketing & Chief Demonstration Officer, Cisco
There’s no doubt that software-defined networking (SDN) is today’s latest buzzword and hot topic. Most IT managers understand that SDN has the potential to improve networks, but SDN often has a different meaning for different people, and that can be very confusing for anyone trying to puzzle out a plan for implementing SDN. Initially, SDN was a term used to explain the concept of splitting the forwarding plane from the control plane, with the added benefit of automation and orchestration. However, recently SDN has become a buzzword attached to a variety of products, which has only muddied the waters.
So how can we clearly define what SDN really is? And is SDN the one solution that will solve many of the CIO’s headaches, or do we need to look beyond SDN? To clarify the many definitions of SDN, it’s helpful to look at SDN as defined in one of these three ways:
• Centralized control, which is especially popular with data center
• Centralized control with lots of distributed intelligence, or
• Software programmability into existing infrastructure, meaning more of an orchestration set.
While SDN can solve many issues with programmability, to achieve the most benefits we need to go beyond SDN alone to encompass a broader view of next-generation infrastructure: orchestration, policy-based automation, virtualization that is broader than SDN, and an application-centric view. CIOs are under pressure to deliver IT value faster and become more agile in response to changing business needs. But to achieve that, IT infrastructure must become less complex, and include more automation, especially policy-based automation. When approached holistically, these capabilities can deliver “Fast IT” – that is, faster, more agile IT enabled by emerging technologies in cloud, automation, virtualization, and big data.
Why Fast IT?
The old way of doing things won’t work anymore for IT professionals. The “application economy” and explosion of connected devices have greatly increased IT complexity--IT managers have to spend too much time just “keeping the lights on.” Most CIOs would like to shift their team’s focus towards more automated processes to free staff time for innovations that support business goals and enable them to respond faster to changing business needs. But to achieve this, CIOs will need to shift their thinking and ultimately, their adoption of infrastructure technologies, to embrace a Fast IT model.
The Road to Fast IT
As with any strategic initiative, the transition to a Fast IT model requires careful planning and change management. In particular, organizations need to develop a plan that encompasses people, organizational processes and technologies. Once this foundational plan is in place, CIOs are then ready to begin the steps of preparing their business for Fast IT and building an organization focused on service delivery.
There are four steps that CIOs and their teams can take to achieve a faster, more intelligent infrastructure. Here’s a roadmap for each step:
Step 1: Virtualize your Compute and Storage Infrastructure Where Applicable
The journey to Fast IT requires tight integration among server, network and storage assets. Traditional servers, storage and network silos have been an impediment to speed and innovation. Standardizing and treating your resources as a single, converged infrastructure pool allows dynamic provisioning and faster time to deployment. Virtualization enables IT to provide services that scale seamlessly across a standardized infrastructure. A single-dashboard approach to deploy and operate virtual and physical data center assets, and wired and wireless network infrastructure is essential to deploy new services across the organization. But your infrastructure must be thoughtfully designed to allow bare-metal provisioning while still providing the benefits of virtualization.
Step 2: Begin using Programmability
In a Software-Defined Architecture, management and control of the networking, storage and data center infrastructure are automated through software to speed up speeding up deployment processes and reducing manual errors.
Programming network behavior is a valuable capability for every organization, especially as they prepare for Big Data, surging video traffic, Bring-Your-Own-Device (BYOD) environments, and cloud computing. Programmability also allows organizations to dynamically modify infrastructure, making it easier to install and run applications.
Step 3: Develop Increasingly Sophisticated Infrastructure Policies
Making your infrastructure application-aware with centralized automation and policy-driven application profiles delivers software flexibility with the added scalability of hardware performance. Through a policy-driven application approach, organizations can optimize applications and experiences in real-time with the assurance that rules for compliance, governance, and service level agreements are met. Instead of programming infrastructure on a box-by-box basis, agility and operational simplicity are realized by defining workflows and policies in advance and across the infrastructure.
But you need to start with simple policies that can leverage controllers in your existing environment and build more and more sophisticated templates that can address specific application deployments (SAP Hana for instance) or compliance requirements (for example, HIPAA in the healthcare industry).
Step 4: Implement a Cloud Strategy
Because one size does not fit all, a combination of public cloud, private cloud and dedicated servers working together delivers the best results in terms of performance, total cost of ownership, investment protection, security and compliance. With the Intercloud– a cloud of clouds– organizations can choose where to place workloads based on technical requirements, performance, price, regulations and user expectations. For example, you may choose to use dedicated infrastructure to launch critical new applications while using a public multi-tenant cloud to handle peak traffic.
Beyond SDN to Fast IT
As you plan, take care to research and select an infrastructure architecture that can take you as far along these steps as possible. After completing these four steps, your infrastructure will be simple, smart, secure and ready to deliver fast, agile IT. And you will have greatly reduced your operating costs. Following this type of roadmap can help IT teams be confident in their decisions and help prepare them for the future. Using one of the steps individually– for example, developing a strategy for the Intercloud– is certainly useful. But implementing all four steps together will reduce costs while offering growth and innovation opportunities to organizations.