John Roese’s Blog CTO, Nortel

Location: Flying to Vancouver

First, let me apologize for being absent from blogging for a bit. I’ve had an unusually busy and complex several weeks. Hopefully, you all found the guest blog from Andy Lippman interesting and had a chance to see some of the recent dialog I have had with other bloggers, such as Om Malik (GigaOM). One of the challenges of blogging when you have a few other full-time jobs is finding quality time to write and, ironically, when you’re not able to write it’s amazing just how much interesting content and dialog emerges, creating a pretty significant backlog of topics. I hope to tackle many of those topics in the coming weeks.

Over the last few weeks, I’ve been in three cities in California, as well as in Boston, Ottawa, and London. I have met with some of our strategic partners (both go-to-market and technology), regulators, investors, media and customers key to our business. One interesting topic that surfaced in many of those dialogs was around whether the use of wireless technology in the enterprise would transform the enterprise networking structure.

As you may recall, last July we announced an initiative called “the unwired enterprise,” where we predicted that by roughly 2010 enterprises would be able to use wireless technology as their primary access network. We said we believed that the combination of innovation beyond 802.11n WiFi, together with the emergence of such 4G wireless technologies as LTE and Mobile WiMAX, would create a model where abundant and functional local and broadband mobile capacity would be real. This would mean the end of the over-wiring of campus LANs, a change in the value chain, and - most importantly - an expectation that our applications and business tools would be built for the mobile world first and primarily (versus today, where they are designed for wired campus LANs and then adapted for wireless).

The interesting thread in the various dialogs, though, was the discussion around “if broadband wireless and 4G emerge, why would anyone need a campus LAN at all? Why not simply move every application and device to the cellular carrier 4G network and eliminate the enterprise network entirely?”

Now before you react too quickly, a pretty strong case exists to make this move. First, if we look forward to LTE and Mobile WiMAX, we can see multi-megabit per second capacity with very low latency and the added bonus of large-scale broadband mobility. Additionally, 4G networks have a very enterprise-centric operating model, where security is designed to decouple the end system from the infrastructure (providing the flexibility required in order to connect your own devices to the network). A 4G network will also have the characteristic of being a “pipe in the sky,” meaning that it is a packet-based IP transport without strict design for any specific applications (just like a corporate LAN or the Internet). All in all, 4G networks will look much more like large-scale WiFi networks or even corporate Ethernet LANs than they will resemble a cellular network of the present or past. That’s pretty appealing and, given the innovation going on in mobile devices (e.g., smart phones and ultra mobile PCs), it seems like this could be a reasonable model for the enterprise.

As good as that sounds, however, the case can also be made that there will still be a solid need for the enterprise to own and operate its own internal network. The reasons for this position are just as compelling.

First is the issue of trust and control. If you own your own network then you are, as they say, master of your own destiny. Given the import of connectedness for business systems, giving up this critical control is unlikely for most CIOs and enterprises. While carrier networks and cellular systems are today a big part of the overall enterprise architecture, they are used in very specific and measured models, usually accompanied by strict service level agreements and the associated costs of such guarantees.

Second, while the 4G world is pretty fast for a cellular system, the capacity innovation inside the enterprise is amazingly rapid. With 802.11n, we are seeing in excess of 300 megs per channel and with some new innovation upwards of 15 or 20 channels per area, for a total of 6 gigs of capacity. Realistically, a system with multiple gigabits of capacity shared among a small number of users anywhere in a campus is possible. This is close to the kind of capacity the wireline gigabit world is offering today and a few orders of magnitude faster than even 4G. Given that capacity mismatch, the allure is there for enterprises to continue to offer local high-capacity networks, be they wired or wireless.

Third, is the issue of intelligence. In most enterprise networks, transport is not just a pipe but a system that allows the CIO to build applications that utilize network intelligence to provide secure and directed services. Consider role-based access control, where a CIO can use the internal network to provide secure access to the company’s employees while also using the same network (in a different policy) to grant visitor or guest access. That kind of flexible security control is easy to do on an internal LAN but would be much more complex in a network provided by a large-scale operator or carrier (where one network would need to support all the roles of all the enterprises sharing it, not just one enterprise’s policy set). In addition to roles and security, the corporate LAN is also increasingly being used to provide location information, to qualify presence and availability to unified communications systems, and to trigger appropriate multimedia interaction that is based on the context of the user and his/her environment (depending on location, state and mobility, for example). This kind of flexible context-based service is a key to the intelligent enterprise experience and would not be easily replicated in a public network system.

Finally, there is the cost issue. Speaking as a former CIO, one thing CIOs like are services and technology that are free once you own the system. Internal voice and data networks fall into that model; cellular and wide area systems do not. The idea of having a networking option that allows for applications and communications to operate at zero cost in even some of the footprint of the enterprise is hard to give up.

So, the question is: “Will innovation in cellular mobile networks render the enterprise LAN unnecessary?” My opinion is “no.” Although we will clearly see dramatically expanded use of 4G networks to support the enterprise, we will simultaneously see a shift in the enterprise to a more mobile WiFi network working with that 4G ecosystem. This combination of both public and private high-capacity, mobile, intelligent wireless access networks is the essence of the unwired enterprise.

This shift to the unwired enterprise will happen over the next several years and with it will be a transformation of the application models used in the enterprise (a shift towards mobility-based applications as the default). It will also mean that the preferred end system will be inherently more mobility centric (laptops, UMPCs, smart phones and, most significantly, multi-network roaming between the 4G and WiFi campus). And, finally, as this new reality emerges the intelligent interaction between infrastructure and applications will create more targeted and intuitive collaboration experiences.

As with any prediction, time will tell. We’ll have to wait and see what actually happens and how everything plays out. The punch line though is that the evolution of the unwired world will change much of what we understand about networks and the new result will be an unprecedented level of interaction between public and private networks and a value that is much higher than ever before.

I welcome your thoughts on this topic.