Technology Trends, Multi-Gigabit Networks, and iWARP

This is the continuing series of special interviews appearing first here in the Canadian IT Managers (CIM) forum from top-ranking business and technology leaders.

In this two part blog series, we continue our discussion with Rick Maule, CEO, President, and Chairman of the Board of NetEffect.

In this blog, I put these final questions to Rick:
_____________________________

Stephen: Which technologies will have the greatest future impact and why?

Rick: For the data center:

Clustering of commodity servers –more computing power for less cost;

Blades –more computing power in a lot less space;

Virtualization – making the most of the resources you buy;

iWARP Ethernet – a single networking fabric to make the most of clustering, blades, and virtualization – meeting the needs of networking, storage, and clustering.

The combination of these technologies will simplify our efforts, increase our flexibility to address change, and reduce our costs.

Stephen: Describe the evolution, challenges, and promise of multi-gigabit technology?

Rick: The information in your data center is a critical resource of your corporation. The growth in demand for this information is not slowing down – currently it is more than doubling year on year – so the infrastructure has to handle more.

This leads to the need for multi-gigabit technology and beyond. To complicate matters, each application type has its own additional unique demands – high bandwidth for networking, high throughput for storage, and low latency for clustering. This disparity in demands has led to a fragmentation of data center communications – unique networking fabrics for each problem. But this path leads to increased complexity and cost. The promise of the new iWARP extensions to Ethernet is a simpler world – one networking fabric that can do it all – without disrupting existing infrastructure and yet continuing the traditional cost benefits of Ethernet.

Stephen: What are the pros/cons of the major competing technologies?

Rick: For each networking challenge, our innovative industry has produced a solution – such as, Fibre Channel for storage and InfiniBand for clustering. But solving the problems one at a time has one main drawback – scale. Solving each problem one at a time guarantees no significant economy of scale because they are not interoperable with each other. So each solution is constrained to the volume needed to solve each specific problem…except for Ethernet. As Ethernet has done in the past, it is doing again – it is embracing change to meet the new challenges while remaining fully compatible with past generations of Ethernet. The iWARP extensions to Ethernet not only solve each of the problems but also leverage Ethernet’s unique economy of scale.

Stephen: Where do you want to take iWARP? Where do you see it heading in five years?

Rick: The iWARP extensions to Ethernet are the next step in the evolution of Ethernet. Near term, they will be used to address areas of the greatest pain in data communications – latency and overhead on servers for the most demanding applications of networking, storage, and clustering. And many of theses applications will reap the benefit transparently – the applications will just run faster with lower use of the CPU. And Ethernet is still Ethernet. The economies of scale will lead to this fully compatible set of extensions finding their usefulness in a wider and wider range of applications. Which will in turn drive increased volumes and lower costs. And once again the question for Ethernet will move from “why” to “why not.” At the right price point, iWARP extended Ethernet will be the only Ethernet … and the cycle will begin again. In five years, we won’t be asking about iWARP – that will be taken for granted; we will be challenging ourselves for what’s next.

Stephen: What is the value proposition offered by iWARP?

Rick: A single solution that meets the needs of the data center– technical, ecosystem, and economic.
Simply put, we need speed without pain. iWARP increases the performance of Ethernet to multi-gigabit speeds while actually and dramatically reducing the resulting load on the server. In doing so, it also gives Ethernet the necessary characteristics for proper support of storage and clustering.

Unlike competitive networking fabrics, it not only solves the technical problem, but also addresses the ecosystem and economics issues. iWARP was designed to extend Ethernet without introducing disruption into the ecosystem of the data center. It is fully compatible with anything you already have. And an IT manager can address his performance and bottleneck problems incrementally – solve the pain where it hurts the most spending as little as needed.

Stephen: Rick, we will continue to follow your future contributions to the business and technology community. Thank you for taking the time to do this interview.

Rick: Thank you for the opportunity.
________________________

I encourage you in the CIM audience to share your thoughts here on these interviews or send me an e-mail at sibaraki@cips.ca.
________________________

Thank you,
Stephen Ibaraki, FCIPS, I.S.P.