While I’ve been writing about Cisco Domain TenSM, I’ve been watching the SDN debate evolve in our industry, and I have to say, I’ve had my concerns. Don’t get me wrong – I personally see SDN as an important and very much required evolution (and note: ‘evolution’ – not ‘revolution’) of the networking industry. Being able to extract more value from the network – through, for example, a consistent and broad network API – I mean, who wouldn’t be excited about that! And especially for us in Cisco, with the largest by far networking installed base, the ability to uncover and exploit additional value for our customers from the network can only be a good thing!
As I say, over the past year or two, I’ve been perturbed about lack of discussion across the industry about the adoption and deployment challenges associated with SDN. There is – bluntly – too much “nirvana” or “marketing promises” out there, too much focus on the end result (e.g. “look at our use case, wow isn’t it great”) without discussion of steps required for a success, and too little discussion on the costs and challenges of the design and implementation of SDN solutions (e.g. “took us X man years + $M of investment”). It’s now time to change the discussion.
I was therefore delighted to see Jim Meltzer’s discussion of the issues he was seeing with his clients regarding SDN.
Much has been made of the emergence of Software Defined Networking and the programmable network. At its core, SDN involves opening up network interfaces in order to make the network programmable and allow for the development of applications. While some of those applications interact directly with the data plane, determining how individual packets are treated, many applications actually involve what can fundamentally be described as management functionality – automation of workflows, reaction to events, closing of control loops. A popular example concerns orchestration, in which resources are allocated and state modified so that collectively a service is provided – in many ways resembling a reincarnation of service provisioning in a new context and under a new name.
Of course, management applications and management interfaces have been around for a long time, so what is really new and different this time? Is SDN simply an exciting new label for a tired old concept? Does SDN obviate the need for traditional management? At the core of these questions are the concepts of programmability and manageability. Read More »
Just back from Varrow Madness 2013 (#VM13) in Durham, NC, a local event for me, and a great opportunity to connect with customers, partners, and cloud knowledge. Here are some highlights.
Rusty Buzhardt (Cisco), Jason Nash (Varrow), Elijah Stukenborg (Chiquita)
Really enjoyed an inspirational keynote by VCE President Frank Hauck and the opportunity to learn about some new technologies that are emerging in the Data Center space. Look for upcoming #EngineersUnplugged episodes for some discussions around NetScaler, Flash Virtualization, and more.
Last week at the RSA Conference in San Francisco, I had the pleasure of speaking to thousands of security professionals about the opportunities and risks associated with using Software Defined Networking (SDN) for security, which will be the underlying fabric of our next generation data centers and networks. SDN-enabled security will provide a better way to secure our most valuable applications, users and data, now and in the future.
Each vendor has a different definition of how the network is changing, and there are many different terms being used, such as software defined data center and software defined storage. Cisco calls this Application Centric Networking, for example, because we are introducing programmable APIs with a focus on distributed control plane intelligence so that applications can get value directly from the network.
It’s obvious why the networking industry is embracing SDN: lower operational costs and the ability to deploy applications and network services in a quicker, more scalable manner. Cloud bursting, which is about flexible compute in the cloud, is another SDN benefit that gives us the ability for applications to interact directly with the network in ways that do not happen today. For example, applications will be able to query the network for location of users to manage Quality of Service and deliver highly targeted content.
So why should the security industry care about SDN? As the threat landscape evolves, the opportunity is to make Security a key application for SDN. We can use SDN to build a Network-based Threat Defense System. I see three key elements to this system:
In this week’s episode of Engineers Unplugged, Brian Gracely (@bgracely) of Virtustream takes on the challenge of explaining the industry’s top buzzword, Software Defined Networking, using doughnuts. Seeing is believing:
Welcome to Engineers Unplugged, where technologists talk to each other the way they know best, with a whiteboard. The rules are simple:
Episodes will publish weekly (or as close to it as we can manage)