Why Deutsche Telekom Created MobiledgeX

 
Sunay_MobiledgeX.png

Sunay Tripathi

CTO, EVP ENGINEERING

It may seem strange or even counter-intuitive that Deutsche Telekom (DT), a mobile operator with a specific cellular telephony franchise in competitive markets, would create MobiledgeX, a new venture chartered to serve any or all mobile operators (not just DT). Here’s the backstory that explains why.

The Catalyst for Change

The world of mobile telephony changed forever a little more than 10 years ago when Steve Jobs announced an open market for iPhone applications. With the next release of iOS, a new application -- the App Store -- would be preinstalled on every iPhone. The App Store worked with a backend application to provide a “store” from which the subscriber could “buy” and install applications (many of the applications are free). Apple designed iOS to isolate applications robustly, and vetted submitted applications carefully before they were available on the store––both important factors in creating an open software market. Before the App Store, the mobile operators controlled what software was installed on their subscribers’ phones, on a device-by-device basis, and used their position in the delivery chain to exact whatever rent the market would bear. With the success of the iPhone, all that was gone forever.

mobile_apps_blog.png

The iPhone and the App Store were wonderful for cellular telephony, and for operators, at least initially. The ease of installing applications fueled subscribers’ interests in music, photos, and then video, all of which created an insatiable demand for more bandwidth. That demand, in turn, drove the 4G/LTE buildout, the need for new handsets, and more data plans. The networks now provide up to 100Mbps cellular service, and end users pay about the same as they did before––an engineering marvel.

But the iPhone App Store left the mobile operator as a commodity bandwidth provider; not the most attractive business proposition. In the last 3-4 years, “edge computing” started to emerge as a Internet/cloud topic, prompting DT to want to learn more, and wondering whether edge computing could be a way by which the mobile operator could play a larger role in mobile data applications.

The idea of edge computing -- providing cloud services nearer to the user or device than today’s cloud services --  is interesting because of the practical impact of network latency and bandwidth, and because of the potential for improving a wide variety of cellularly connected “client/server” applications. The Internet is a miracle of connectivity, but TCP/IP was designed to be open, easy to manage, and capable of being extended for unanticipated uses -- all of which it has done superbly as today’s Internet demonstrates -- but as a result of those design choices, has made the Internet weak when it comes to performance and service assurance.

Edge computing can’t be used to restore what was lost with the App Store, but it can be used to add real value to mobile devices and connectivity. For almost all high-performance applications that require device and application interaction (e.g., watching Netflix on your phone, playing a multiplayer game), the less the intervening network, the better; which is exactly what moving resources to the “edge” of the cloud can do. Today’s cloud computing began in large, centralized data centers for good reasons. But in the last decade, cloud computing has become highly automated (depending less or very little on human management) and now can be distributed much more broadly because most cloud data centers can be left unmanned, even in “lights out” locations. Edge computing takes advantage of that to improve application performance by improving the network connectivity between the edge and the device.

It really is as simple as that.

The Edge Opportunity

MobiledgeX doesn’t claim to know all the valuable applications yet, but that doesn’t bother us. If edge computing is like most disruptive innovations in the past, the value will be discovered by early users, not by system’s architects or vendors. MobiledgeX is just starting the process of building out the platform, talking to likely application owners and developers, and creating a marketplace: the equivalent of the App Store, but for edge applications and services. We think many of the most valuable edge uses will be discovered by that community, not just by MobiledgeX.

There is a catch, of course. Applications that assume some level of network performance (latency, bandwidth, packet loss) often don’t improve that much when the network connection is improved. But, that’s not to say that the applications can’t be improved with better network performance. It just requires factoring the application to assume better connectivity. There is every reason to be optimistic: usability experts know that today’s network applications, with response time measured in a few seconds, are too slow to provide ideal usability, and there is a lot of room for valuable improvement. And then there are a lot of exciting new applications (e.g., AI, machine learning, autonomous systems) that require better connectivity to be built at all

MobiledgeX sees many categories of likely benefit (read Finding the Edge at Scale and Edge as an Opportunity), but we believe that much of the specific innovation will come from the early developers and application owners, who see incremental value by incorporating our platform and service through the marketplace we’re creating. Our task, for now, is to evangelize the opportunity.

Leveraging and managing those mobile operator resources, we can provide on-demand, pay-as-you-go virtual computer and service resources near the edge, all with managed and predictable network connectivity to a cellular smartphone or IoT device.

There is a 1.7 trillion dollar global cellular infrastructure that doesn’t really participate in modern cloud computing except to provide raw wireless bandwidth. Unlike many of the Internet network service providers, the mobile operators have a lot of compute power right near the mobile radio edge (as is required to operate mobile telephony and messaging services). Leveraging and managing those mobile operator resources, we can provide on-demand, pay-as-you-go virtual computer and service resources near the edge, all with managed and predictable network connectivity to a cellular smartphone or IoT device. That can be used to extend cloud applications closer to the edge, improving performance and usability; it can be used to enable new IoT applications that leverage AI and machine learning, and depend on that network performance; it can be used to augment or implement new devices and applications based on VR and AR; and it can be used to enable new levels of collaboration among geographically clustered users, like new multi-user games.

Edge Computing Collaboration

As DT studied edge computing, it found many ways in which it could enable greater participation in the cloud, and that warranted further exploration. It saw that many of those opportunities (augmented A/R for example) would require support from multiple mobile operators, not just DT. This was the reason that DT created MobiledgeX as an independent company, chartered to serve all mobile operators, not just Deutsche Telekom.

“Many of the most exciting opportunities we saw were global,” said Tomasz Gerszberg, Senior Vice President of Business Operations, Product Innovation at Deutsche Telekom. “Deutsche Telekom is a large enterprise but we’re not global in that sense. We also saw that these opportunities depended on the collaboration of many operators. For both reasons, it was clear that MobiledgeX would be much more likely to succeed if chartered as an independent company, free to work with many operators, rather than a DT subsidiary.”

Free to serve many mobile operators, MobiledgeX exists to improve the business opportunity for all––while at the same time enabling much greater optimization of the 1.7 trillion dollars of global cellular infrastructure.