The concept of converged IT infrastructure is not new to most businesses. Enterprise CIOs have been dealing with the challenge of getting servers, storage, networks and applications to work together for a long time. So why have another discussion on Converged Infrastructure now? The answer lies in the the need of digital initiatives and businesses to transform IT to an agile internal service provider that is able to offer computing platforms, storage and network services on demand, which is making CIOs look for fresh alternatives that could meet the new business realities while optimizing the total cost of IT ownership.
Gartner refers to Converged Infrastructure as Integrated Systems. In a report titled Magic Quadrant for Integrated Systems published on 16 June 2014, Gartner defines integrated systems as combinations of server, storage and network infrastructures, sold with management software that facilitates the provisioning and management of the combined unit. The market for integrated systems was estimated at over $6 billion in 2014, a 50 percent increase over 2013, and is projected to grow y-o-y at similar pace in the foreseeable future.
Network infrastructure is an important building block in converged infrastructure. Cloud providers that operate hyperscale data centers were the first to realize the importance of having a flexible network fabric where network services could be turned on/off and configured/re-configured dynamically with shifting application workloads and changing traffic patterns. With large scale adoption of virtualization and cloud computing within the enterprise, the same concepts are now being adopted by Enterprise IT to offer bandwidth and network services for distributed dynamically provisioned workloads.
Rapid advancements in network virtualization and software defined networking technologies during the last five years have seen mature IT organizations, financial institutions and large enterprises reinvent the way in which networks are built; and services such as mobility, unified communications, security and business continuity/disaster recovery are made available to business users across enterprise HQ, branch office and remote locations.
Network design can no longer be done in silos for the datacenter, for WAN, or for a specific service like Unified Communications. Today’s distributed applications exchange data and utilize services hosted across multiple enterprise locations, cloud providers and/or partner locations to meet different business needs. Therefore, the new network is not about network connectivity alone; rather, it interacts with business applications and provides a host of network services for distributed applications to work together to offer a business solution.
According to IDC, inadequate networking architectures are causing operational, performance and scaling issues for many of today’s enterprises. Therefore, CIOs need to pay greater attention to network architecture and design aspects, keeping in mind the changing business priorities and the organization’s appetite for leveraging technology to deliver a better customer experience, streamline business processes and improve agility in responding to business challenges.
In the converged network transformation journey, network design has a significant impact on all IT purchases including servers, NICs, HBAs, storage, hypervisor(s), orchestration tools, etc., apart from switches, routers, security appliances, L4-L7 functions, unified communications solutions, etc. that have been traditionally considered within the realm of networking.
Here are few key pointers to remember when implementing converged network infrastructure solutions:
# Build your own Reference Architecture: Pre-configured converged infrastructure solutions offer a quick capacity fix but do not necessarily help enterprises transform their infrastructure. Enterprises should spend time and effort in evaluating open reference architectures and make their own infrastructure blueprint to support present and future business applications.
# Select the right Management & Orchestration Platform: Migration to converged infrastructure needs careful evaluation and selection of integrated provisioning and management tools that go beyond DevOps or network monitoring and troubleshooting, to provide a unified “application + infrastructure” orchestration and management platform. Tools that build on open source platforms like OpenStack, and support the addition of new features/functions offer greater investment protection.
# Aggressively Develop New Skills: Migration to converged infrastructure requires IT teams to develop new skills to be able to provision, monitor and troubleshoot the new collapsed infrastructure, often with multiple layers of abstraction. Enterprises should consider partnering with organizations that offer expert resources to coach and train internal IT staff to bridge skill gaps during transition.
# Redraw the RACI/RASCI matrix: IT support processes need to be rewritten and the Responsibility Assignment Matrix for IT support staff needs to be redrawn for converged infrastructure support.
# Encourage Innovation: Apart from ROI and TCO gains, one of the important considerations in converged infrastructure purchases is the ability of the chosen building blocks to support innovation in a service-oriented world. The new converged network should support programmability and flexibility to add new services over IP to keep pace with the growing distributed application infrastructure. Indeed, the goal of SDN and NFV is to facilitate greater network programmability, and open the door for innovation which was not possible with fixed function network elements earlier.
By Mrinmoy Purkayastha, VP, Alten Calsoft Labs