All of a sudden, data centre networking has gone from ho-hum to hot.
The status quo — that hierarchical, three-tier network that has dominated the data centre since the late 1990s – is being threatened, as enterprises move toward a virtualised, service-oriented, scaled-out, converged infrastructure.
The question facing enterprise IT architects and planners today is whether to stick with three tiers or to flatten out the network in order to improve performance and ultimately save money.
From the vendor perspective, companies like Brocade, HP and Juniper Networks are pushing hard for the flat network. After all, if enterprises decide to go flat, they will need to buy new data centre switches and that opens up an opportunity to displace Cisco, which for years has been all but synonymous with the data centre network.
“Even companies with Cisco Catalyst switches have to re-qualify their data centres and, to go into flat network mode, they'll have to look at all brand new equipment. This is a huge opportunity for others to come and take the data centre network away from Cisco” says Andre Kindness, senior analyst for enterprise networking at Forrester Research.
One of the underlying realities of clouds or shared resource pools is that larger clouds and more dynamic clouds enable greater efficiency and flexibility. Unfortunately the desire for greater scale and dynamism places significant pressure on the data centre network. It is the network or networks that interconnect the resources and enable the creation of a cloud. No network, no cloud. Unfortunately, networks are also the greatest impediment to success, as most networks today are not architected to support the scale or dynamism demanded by the modern data centre. Only by re-architecting the data centre network can companies realise the full benefits of virtualisation and cloud computing.
The typical architecture for a data centre Ethernet network employs a tree structure, arrayed in three layers of switching, fanning out from the core. This design was derived from the architecture of the LAN and was adopted in data centres approximately 10 to 15 years ago when Ethernet displaced SNA, token ring and DEC-Net. As the data centre evolved over the last 10 years, the network began to exhibit shortcomings, including suboptimal performance, inherent inefficiency, and excessive complexity. None of these shortcomings is trivial, but it is the complexity that truly stands in the way of the virtualised data centre.
The complexity arises from the fundamental architecture of the tree structure. Networks are comprised of multiple, autonomous switching devices that cooperate through shared protocols. Managing a network entails managing not only the switches but also the interactions between the switches.
“Many organisations have taken on server, application, and data centre consolidations to reduce costs and increase returns on their IT investments. To continue their streamlining efforts, many organisations are also considering cloud computing as a new infrastructure model to create dynamic scalable resource pools. However, many data centre networks in operation today were designed for 5 to 10 year old traffic patterns and are not optimal for the current application and traffic flow environment. These designs typically result in lower performance, unnecessary complexity, and higher costs. While most organisations accept the need for change, the vast majority don’t want to disrupt their operational production data centre to begin the transition to a simpler, high-performance network design,” says Tarek Abbas, Systems Engineering Director, Juniper Networks, Middle East and Africa.
Brocade is another major vendor pushing for a flat data centre network. “Data centre networks rely on Ethernet. Over the decades, Ethernet has evolved as new types of architectures emerged. Today, data centre networks carry traffic for a diverse set of applications including client/server, Web services, unified communications, virtual machines, and storage—each with different traffic patterns and network service requirements. For example, when Ethernet carries block storage traffic, it places stringent demands on the network including lossless packet delivery, deterministic latency and high bandwidth.
Today’s infrastructure is not sufficient to handle new applications, and not flexible enough for a world where applications are mobile. It’s clear that data centre networks need to be upgraded over time to increase performance, reduce latency and eliminate downtime. They also need to be designed specifically to support highly virtualised and cloud-optimised data centres,” says Ali Ahmar,Regional Sales Manager, MENA, Brocade Communications.
Avaya, which has acquired Nortel, has a new data networking architecture that it says will help enterprises reap the benefits of virtualisation in a more simplified and cost-effective manner. Called Avaya Virtual Enterprise Network Architecture (Avaya VENA), the company says the goal of the new architecture is to allow organisations to more easily optimise business applications and service deployments in and between data centres and campuses.
“Simply put, the goal is to make the Core of the network transparent. By evolving Data Centre and the Campus Core infrastructures into one that seamlessly integrates services – and by this we typically mean applications delivered by virtualised computing systems – availability, performance, and efficiency can be dramatically improved. Simplified service orchestration is possible when the network facilitates an easy and effective, yet highly granular, mapping capability,” says Maan Al-Shakarchi, Senior Sales Manager, Data Network Solutions – Avaya, Emerging Markets
Cisco, of course, has next-generation plans of its own, built around its Nexus switches. “The architecture of the Data centre is changing dramatically. The next generation Data centre is transforming from business-centered physical location into highly integrated and virtualised environment allowing our customers to scale their resources, their information and most importantly their business. The explosion of video; real-time, interactive applications; and communications and collaboration technologies has created more network-centric applications, requiring a new data centre platform,” says Ammar Halabi.
On the actual design of the data centre, Cisco solution allows for a 3 tier architecture (3 physical layers) as well as a flat architecture (1 physical layer virtualised into 3). “The challenge is not whether to have 3 physical layers or 1. The challenge is in understanding the exact role of each layer and building an architecture tailored to the specific needs. Our vision is the evolution of the data centre to a more consolidated, virtualised and automated environment that lays the foundation for cloud computing,” says Halabi.
From the IT side, the allure of the flat network is being able to deploy products purpose-built for automation, convergence and virtualisation. The promise is a much simplified, fabric-based architecture from which enterprises will enjoy dramatic performance improvements as well as streamlined operational chores, and expenses.
“The limitations of legacy networks are many. They include unreliability and unpredictability, difficulty in achieving effective scale, unavoidable complexity, inefficient resource utilisation, delays in provisioning new services and slow time-to-service, and difficulty in maintaining separation of traffic. These limitations result in an inability to decouple networking considerations from service delivery, introducing a genuine element of constraint,” says Al-Shakarchi.
He adds that to support the transition to a multi-dimensional environment, the underlying network also needs to change. Provisioning needs to be simpler, and availability and performance need to scale seamlessly.
Halabi offers another perspective on why companies have to re-think their data centre network strategy. “We have to understand the operational challenges and business pressures that organisations are facing today. Almost all the CIO’s we’ve had the privilege to work with over the past few years concur that around 80% of their budgets is being consumed just to keep the lights on. Energy cost and availability is a major challenge. Asset utilisation is around 7% because of the way data centres have been built over time and the disparate islands of resources that does not enable efficient and shared usage of those resources rendering them drastically underutilised while you have to keep fully powered up and maintained.”
Those silos in data centres made it increasingly complex and lengthy to provision new services which dramatically affected business agility and time-to-market. Information needs to be in real-time, and collaborative business applications are putting a huge demand on data centres, and more often than ever, we needed to respond to the economic uncertainties we are living in. All that made status quo not an option anymore. The industry not only needed to respond to those challenges and pressures, but also find a way for CIO’s to free some of their budgets for innovative projects rather than spending 80% of an already shrunk budget on operations and maintenance, he adds.
Price and performance rule
To flatten or not needs to come down to a price/performance evaluation. Flattening the network will require a major investment in new equipment. Network executives must be able to tell a good financial story about how flattening the network will increase revenues or reduce costs outside of the data centre. That'll be hard for the average enterprise to do. Data centres are changing, but the decision points for network managers remain the same. If a two-tier network can give you great performance at a better price than a three-tier design, it wins because theoretically it should have lower latency and even better reliability too.
The decision doesn't need hurried, but a careful study of options and setting of expectations, Kindness says.
“This isn't going to happen for two to five years out and even then not as one big changeover at once,” he says. “This is going to be an evolution, and no one should expect their vendors – server, storage or networking – to enable this to happen automatically.”
Whether it happens tomorrow or not, what is for sure is that to realise the full benefits of virtualisation, and enable more scalable, more dynamic resource pools, companies need to re-architect the data centre network with an eye towards eliminating its inherent complexity. By converging on a single network, flattening the network, reducing the number of separate switches and switch interactions, leveraging the single control plane of a network fabric, and leveraging state-of-the-art automation, one can simplify the network, uncapping the potential of the modern data centre.