One of the hottest topics in networking at the moment – if not the whole of IT – is the concept of Software Defined Networking (SDN). A recent IDC (International Data Corporation) report, Technology Assessment: The Impact of OpenFlow on Data Center Network Architectures, predicts that the OpenFlow and SDN market will grow to almost $2 billion by 2016. So what is SDN and what is all the noise about?
The move towards highly virtualised environments and environments where rapid changes to service delivery occur has created challenges for traditional switched Ethernet networks. Software Defined Networking aims to overcome these limitations by separating the management and control planes of the switch and moving these into a separate controller. This leaves the network hardware – the switches and routers – to do what they do best, forward traffic.
One of the aims of SDN is to deliver greater agility and flexibility to environments where this is desperately needed – namely telcos, service providers and data centres. This would enable services to be rapidly deployed and would simplify the migration of environments from one location to another. What is less clear is what benefit, if any, enterprise organisations and other institutions, such as schools, will achieve from the deployment of SDN.
In moving the intelligence from individual network devices to a centralised controller, there is the belief by many that this will commoditise networking hardware and result in lower cost networks. It is true that much of the value within today’s networking devices can be found in the software that runs on these devices and not the hardware itself. As a result of removing the software from the device, it is highly likely that SDN will decrease the cost of hardware – however, hardware is only one part of the network. In moving the intelligence to a centralised controller, schools will need to make a significant investment in the controller and required applications to manage and operate an SDN network. In addition, due to the complexity of SDN, it will be necessary to hire or contract additional skills to manage or oversee an SDN environment. These factors are likely to deliver increased agility and flexibility, but will come at a cost as a result of additional knowledge and skills required to oversee a school’s network.
While telcos, service providers and data centres are grappling with challenges associated with scalability, flexibility and agility, schools are dealing with a completely different set of challenges. Sit down and speak to an IT manager within a school and they will talk about complexity and a desire to simplify their IT environment. They will talk about the need to deliver outcomes in the most efficient manner possible. SDN is unlikely to help schools with these challenges.
You could argue that some form of Network Management System (NMS) is one way to simplify the management of an IT environment. The reality is that in order to manage the greatest number of devices, including servers, storage and networks, users have to sacrifice some extended functionality. An example is the ability to configure QoS or ACLs across a network of switches, which is challenging if not impossible using third-party NMS platforms. The end result is that schools that do deploy an NMS, typically use it as a monitoring tool rather than a management tool. This means they have to resort to managing discrete devices independently or use yet another platform for managing devices, like a network of switches, which again increases the cost and complexity.
While the primary benefits of SDN are increased agility or flexibility, which are primary requirements in large data centre environments, schools are grappling with reduced complexity and decreasing the total cost of ownership of their infrastructure. SDN is unlikely to help in overcoming the challenges of enterprise organisations so it is necessary to look at other ways to overcome these challenges.
What if there was another way?
While switches and routers are perceived to be expensive, the reality is that the majority of the cost of operating a network is in the labour and skills to deploy and manage the network over time. One customer recently commented that even if you gave him (at no cost) switches to replace his network, he could not do it as he did not have the budget to perform the upgrade and manage the expanded network over time. And if we consider the complexity that schools see within their infrastructure, this challenge is largely associated with making the environment (or network) operate as a single unified entity.
These challenges can all be overcome by changing the management plane of the network associated with how devices are managed. Imagine a network where, when a new device was connected, it was automatically configured by the network itself. Or a network where firmware and configuration is automatically backed up so that should a device fail, all that is required is to replace it with an identical device and it would be automatically configured and operational. Or making a configuration change to a network was as simple as entering a single command once and it would be automatically pushed out to all devices across the network.
This type of functionality would go a long way toward simplifying the management of networks, reducing the amount of time and effort required to manage the network, driving down the Total Cost of Ownership, and resulting in greater uptime and end user satisfaction as the chances of errors or omissions being made during changes are greatly reduced.
When you are considering SDN or looking to upgrade your infrastructure, here are four things to consider:
- Develop a list of outcomes you expect from your technology upgrade.
- Assess suitable technologies that will achieve these outcomes (ensure outcomes are driving the technology you chose, rather than the other way round).
- Develop a deployment plan to migrate from you current environment to your new environment (ensure you have a plan for rolling back if things do not work out as planned).
- Measure your expected outcomes against actual outcomes:
- Did you achieve what was expected?
- If not, why not? What would need to change in order for the upgrade to be successful?
In many instances, schools become fixated on a specific technology which fails to deliver because, as great as the technology may be, it was not suitable for their environment. Always ensure your business outcomes are driving the technology you use and you are more likely to succeed in achieving those outcomes.
Scott Penno is a Principal Evangelist and Asia Pacific Regional Marketing Manager for Allied Telesis, an information and technology services provider to the education sector. He can be contact at firstname.lastname@example.org.
Latest posts by Education Technology Solutions (see all)
- PD 14 Performance - October 24, 2018
- Record Number of Competitors Team up to Reverse the Threat - October 10, 2018
- PD 14 Touchbar - October 4, 2018