Sunday, April 25, 2010

Enabling Innovation with next generation wireless 5G Internet + clouds - technical details

[As I have blogged several times I believe there is a huge potential to create a platform for innovation through the integration of next generation wireless Internet with cloud mobile applications. I recently have a talk on this subject at an OttawaU WiSense seminar -

The new 5G wireless networks have 2 major features – they can be powered solely by renewable energy and enable tight integration with clouds. The WiFi nodes can be powered solely by renewable energy using power over Ethernet or 400 Hz multiplex power on existing 60Hz power lines. With overlapping coverage and the use of 3G/4G as a backup this type of arrangement is quite feasible. An early example of this architecture is the Green Star Network ( The Internet applications that will run over mobile devices will largely be supported by clouds and deeply distributed content networks as the consumer mobile devices will not have the computing power to support advanced applications, such as using the mobile device as a sensor. I have provided some technical details on what a 5G network would look like with some application examples are given in my talk above. Also see my paper on the integration of 5G networks and clouds at

Integrating Wifi with 3G/4G networks is not a new idea. But in the past Wifi networks were seen as a second cousin to the carrier’s mobile network, and even if the Wfi network was for local service the end to end connection was operated and controlled by the carrier. With 5G networks the exact opposite arrangement is possible where the enterprise WiFi network is the controlling network and only uses the 3G/4G as back up and fill-in when there is no WiFi service. Several companies are already in this space such as BelAir Networks ( and Stoke (

I believe R&E networks can be early leaders in this market as they have high capacity backbone networks and their clients such as universities and schools have extensive WiFi networks. Some institutions have deployed an early progenitor of a 5G network with tools like Eduroam. In countries with competitive telecom markets there are now a number of virtual mobile wholesale service providers who would be willing to provide the backup 3G/4G network. A good example is Harbingers proposed global LTE wholesale LTE service

Analysts have always wondered why Cisco has not entered the 3G/4G market. They clearly have dominated so many other aspects of telecom and Internet. The one obvious missing market segment in Cisco’s strategy is mobile. I suspect that Cisco is also planning a Wifi over 3G/4G strategy as well. It makes sense. Data networks are a completely different beast than voice networks. Cisco has never dabbled in voice network technology. No wonder. Today’s mobile networks are very complex using arcane and convoluted voice protocols. As more and more mobile traffic is Internet based the need for a new, more data efficient, far less complex wireless Internet architecture is required. I would not be surprised we see some major announcements from Cisco and Google in this space later this year. Cisco brings its Wifi network expertise to the table and Google brings the cloud infrastructure plus the Android to support the applications.

Japan is already moving in this direction as noted in an e-mail exchange with Herman Wagter on Gordon Cook’s ARCH-ECON list:

“The demand for ubiquitous broadband experience in Japan and its effects is a lesson we should learn from and prepare for.

I understand Softbank is subsidizing customers to get as many picocells connected to home-broadband-wirelines as possible, to offload as much traffic as possible to residential lines. This strategy is much faster and flexible (add where the demand is) than building more masts. In urban areas, where the demand is the highest it is hard to add cell towers (negotiations, permits, costs) and easy to add picocells to residential lines.

In a couple of blogposts I recently have elaborated on a realistic and cheap architecture, using existing technology, to support various kinds of use of the broadband line other than Internet access, like these picocells. Broadband as a utility, taking virtualization one step further.”

Finally here some excerpts from a blog about the innovation potential of clouds

Startups, cloud computing, and the freedom to innovate
Cloud computing is grabbing a lot of headlines these days. As we have seen with SOA in the past, there is a lot of confusion of what cloud computing is, a lot of resistance to change, and a lot of vendors repackaging their products and calling it cloud-enabled. While many analysts, vendors, journalists, and big companies argue back and forth about semantics, economic models, and viability of cloud computing, startups are innovating and deploying in the cloud at warp speed for a fraction of the cost. This begs the question, “Can large organizations keep up with the pace of change and innovation that we are seeing from startups?”

Innovate or Die
Unlike large well established companies, startups don’t have the time or money to debate the merits of cloud computing. In fact, a startup will have a hard time getting funded if they choose to build data centers, unless building data centers is their core competency. Startups are looking for two things: Speed to market and keeping the burn rate to a minimum. Cloud computing provides both. Speed to market is accomplished by eliminating long procurement cycles for hardware and software, outsourcing various management and security functions to the cloud service providers, and the automation of scaling up and down resources as needed. The low burn rate can be achieved by not assuming all of the costs of physical data centers (cooling, rent, labor, etc.), only paying for the resources you use, and freeing up resources to work on core business functions.
I happen to be a CTO of a startup. For us, without cloud computing, we would not even be in business. We are a retail technology company that aggregates digital coupons from numerous content providers and automatically redeems these coupons in real time at the point of sale when customers shop. To provide this service, we need to have highly scalable, reliable, and secure infrastructure in multiple locations across the nation and eventually across the globe. The amount of capital required to build these datacenters ourselves and hire the staff to manage them is at least ten times the amount we are spending to build our 100% cloud based platform. There are a hand full of large companies who own the paper coupon industry. You would think that they would easily be the leaders in the digital coupon industry. These highly successful companies are so bogged down in legacy systems and have so much invested in on-premise data centers that they just cannot move fast enough and build the new digital solutions cheap enough to compete with a handful of startups that are racing to sign up all the retailers for this service.
Oh the irony of it all! The bigger companies have a ton of talent, well established data centers and best practices, and lots of capital. Yet the cash strapped startups are able to innovate faster, cheaper, and produce legacy free solutions that are designed specifically to address a new opportunity driven by increased mobile usage and a surge in the redemption rates of both web and mobile coupons due to economic pressures.
My story is just one use case where we see startups grabbing accounts that used to be a honey pot for larger organizations. Take a look at the innovation coming out of the medical, education, home health services, and social networking areas to name a few and you will see many smaller, newer companies providing superior products and services at lower cost (or free) and quicker to market. While bigger companies are trying to change their cultures to be more agile, to do “more with less”, and to better align business and IT, good startups just focus on delivery as a means of survival.
Legacy systems and company culture can be boat anchors
Startups get to start with a blank sheet of paper and design solutions to specifically take advantage of cloud computing whether they leverage SaaS, PaaS, or IaaS services or a combination of all three. For large companies, the shift to the cloud is a much tougher undertaking. First, someone has to sell the concept of cloud computing to senior management to secure funding to undertake a cloud based initiative. Second, most companies have years of legacy systems to deal with. Most, if not all of these systems were never designed to be deployed or to integrate with systems deployed outside of an on-premise data center. Often the risk/reward for reengineering existing systems to take advantage of the cloud is not economically feasible and has limited value for the end users. If it is not broke don’t fix it! Smarter companies will start new products and services in the cloud. This approach makes more sense but there are still issues like internal resistance to change, skill gaps, outdated processes/best practices, and a host of organizational challenges that can get in the way. Like we witnessed with SOA, organization change management is a critical element for successfully implementing any disruptive technology. Resistance to change and communication silos can and will kill these types of initiatives. Startups don’t have these issues, or at least they shouldn’t. Startups define their culture from inception. The culture for most startups is entrepreneurial by nature. The focus is on speed, low cost, results.
Large companies also have tons of assets that are depreciating on the books and armies of people trained on how to manage stuff on-site. Many of these companies want the benefits of the cloud without given up control that they are used to having. This often leads them down an ill advised path to build private clouds within their datacenter. To make matters worse, some even use the same technology partners that supply their on-premise servers without giving the proper evaluation to the thought leading vendors in this space. When you see people arguing about the economics of the cloud, this is why. The cloud is economically feasible when you do not procure and manage the infrastructure on-site. With private clouds, you give up much of the benefits of cloud computing in return for control. Hybrid clouds offer the best of both worlds but even hybrids add a layer of complexity and manageability that may drive costs higher than desired. We see that startups are leveraging the public cloud for almost everything. There are a few exceptions where due to customer demands, certain data are kept at the customer site or in a hosted or private cloud, but that is the exception not the norm.
The Zapthink Take
Startups will continue to innovate and leverage cloud computing as a competitive advantage while large, well established companies will test the waters with non-mission critical solutions first. Large companies will not be able to deliver at the speed of startups due to legacy systems and organizational issues, thus conceding to startups for certain business opportunities. Our advice is that larger companies create a separate cloud team that is not bound by the constraints of the existing organization and let them operate as a startup. Larger companies should also consider funding external startups that are working on products and services that fit into their portfolio. Finally, large companies should also have their merger and acquisition department actively looking for promising startups for strategic partnerships, acquisitions, or even buy to kill type strategies. This strategy allows larger companies to focus on their core business while shifting the risks of failed cloud executions to the startup companies.

twitter: BillStArnaud
skype: Pocketpro