Ethernet networks have evolved dramatically in capabilities and scale over the past decade. Networks have now become enablers of business-critical applications rather than simply communications infrastructure. Companies are demanding things of their networks that Ethernet was never conceived to provide. Systems are demanding predictable performance in highly complex environments to provide service quality.
Critical applications and communications are reliant on the network, demanding service-provider levels of capability in the enterprise. And the network is being called on to provide support for emerging technologies such as fabric convergence and virtualisation across computing, storage and security.
At the edge of the network, wireless access is becoming an expected feature. Initially driven by the uptake of notebooks, wireless was tolerated as patchy, slow and unreliable. Increasing usage has driven user expectations ever upwards. New generations of smart mobile devices are being adopted that are only able to connect wirelessly. They are being used to access critical applications and communications systems and need to be catered for and secured. This blended wireless edge needs to be managed and secured in an equivalent or even more capable manner as the traditional wired access network.
The network edge also has to support increasingly distributed networks and systems, with branch offices having a combination of local systems and also needing access to centralised services. Being able to effectively monitor service quality will be vital, as will optimising traffic between locations to reduce costs and allow more applications and users to be supported without having to re-invent connectivity or continuously add network bandwidth.
In the datacentre, networking has become a complex, multi-protocol and multi-tier beast. Managing all these adds additional complexity and overheads, increasing both the capital costs to acquire the networking equipment, and vastly increasing the operating costs of managing the network. Getting the beast under control will require more than improvements in management integration, although this will be a valuable improvement and will be more easily implemented than a wholesale rethink of the network.
But a rethink is exactly what is needed to prepare for future requirements that will be placed on networking. Virtualisation is emerging from lab rat to production tool, encompassing computing, storage, networking and security. Moving from a topology and technology-based architecture to a unified and virtualised one is key. Today, the management of secure network connectivity in dynamic virtual infrastructures poses many challenges, some of which are holding back the deployment of more sophisticated solutions offering improved business value.
To achieve this, network consolidation is inevitable. It will take time, and there will be learning experiences before it becomes mainstream. But moving non-Ethernet networks onto the emerging Converged Enhanced Ethernet will greatly aid the flexibility and uniformity of connectivity that virtualisation needs. It should also make administration more straightforward, more cost-effective and thereby more secure and resilient.
But it does not stop there. Reducing the need for skilled network specialists to monitor and manage the network will be vital. If convergence is all that happens, managing performance and service quality will become almost impossible to achieve at scale and cost much more as demands on the network become greater. For every increase in the level of interaction in the network where traffic must be processed and policy applied, uncertainty in performance – particularly latency and jitter – increases, often at an exponential rate.
Obtaining that last 20 per cent of performance often results in a huge jump in costs. Removing tiers from the network can help remove uncertainties and allow more predictable performance. This in turn can allow the use of less equipment and lessen management requirements, both for configuration and ongoing operations. Managing fewer boxes should also help limit risk exposure as routine operations are simplified.
The network will be getting smarter too. It will be an active participant in the new virtual world, aiding and abetting virtual systems as they pop up, move about and disappear by migrating policy, security and identity in context as they move through the network fabric. Automation is the key, as manual intervention is anathema to the innovations, and cost of operations, that virtualisation brings.
Getting all these elements in place will be a big task, as much of the technology is still immature. Greenfield sites and those companies that rely on networking for core competitive advantage are best suited to consider ways to implement them today, even if the solution is sub-optimal compared to the vision. So long as the implementation is flexible to allow continual improvement, the payback should be there.
But for most companies, the pragmatic approach should apply. Different elements will mature more quickly, volumes will increase and prices will drop. Building in the new capabilities gradually in the areas that make most sense will be the preferred approach. It is also the only approach that will be economical and service delivery friendly.
Content Contributors: Andrew Buss & Tony Lock
Tony is an IT operations guru. As an ex-IT manager with an insatiable thirst for knowledge, his extensive vendor briefing agenda makes him one of the most well informed analysts in the industry, particularly on the diversity of solutions and approaches available to tackle key operational requirements. If you are a vendor talking about a new offering, be very careful about describing it to Tony as ‘unique’, because if it isn’t, he’ll probably know.
Have You Read This?
The pandemic and productivity: a Covid-19 conundrum
Lifecycle Management of HCI Systems
Modern Data Protection for HCI
Manage your data, not just your storage
Analytics-Driven Storage Management
Make the camera work for you, not against you
The role of machine learning and automation in storage