Edge as an Opportunity
September 16th, 2018
I was invited to present to the GSMA Board of Directors, here is what I shared:
At Mobile World Congress America we started to talk about MobiledgeX (MeX), a Silicon Valley, Deutsche Telekom (DT)-founded startup that began operations in January. I’m going to give a very high-level explanation of what we’re doing here, and why. You can expect more details to emerge over the coming months.
Our goal is to provide a seamless experience so everyone wins and no one has to abruptly change how they do business.
Stated most simply, our goal is to better integrate the global cellular infrastructure into the modern cloud and Internet, and by doing so deliver benefits to the collective stakeholders. The problem is difficult because these two worlds are very different in all dimensions – engineering, operation and business model. Our goal is to provide a seamless experience so everyone wins and no one has to abruptly change how they do business.
MobiledgeX was created by DT after an intensive internal study about concepts of “edge” computing (putting computation resources nearer to the user of the Internet). The research came up with a lot of interesting ideas, many of which would clearly be better if incorporated by mobile operators –not just DT. MobiledgeX was created shortly after as an independent entity able to form partnerships broadly and to create the needed software. We were not the only result of this effort, it also created tooz technologies Inc., the smart glasses company co-founded with Zeiss.
A Little Background
To understand the opportunity, it’s worth talking a little about how we got here. A lot of the value of the cloud and Internet have to do with communications and collaboration in many forms. That all started with Telephony – one of the seminal inventions of the last 150 years. Telephony systems began as national monopolies and grew and evolved as elegant examples of vertical integration and optimization – a service designed and built from the wires up. Mobile telephony emerged from the same history a century later. Today’s global cellular services are remarkable because they just work (turn on your phone and dial a number) and they work on a global basis through a federation of independent providers. How that was accomplished is fascinating and largely unknown story. The result speaks for itself.
The Internet is “something entirely different” – as Monty Python might say. The differences are important because they create the challenge and opportunities that MobiledgeX is addressing.
The Internet was explicitly designed to be very different from the telephony network. The telephone network was purpose engineered and built for one task: to provide optimal voice telephone connections.
In contrast, the Internet was designed on the assumption that its use would change over time; which has happened beyond anyone’s wildest dreams. As a result, nothing that was application or use specific was designed in – all of that was to be done over the top using a simple, best effort shared transmission structure. Where the telephony network was centralized and carefully managed, the Internet is completely distributed without central management. Whereas interconnection between telephony systems was controlled by international regulators, most of the Internet peering was informal, best-effort and typically without any business contract. The global cellular infrastructure and the Internet are both digital networks, but that’s about all they have in common.
So these two worlds – mobile telephony and the Internet – evolved quite independently until today’s smartphone was introduced only about a decade ago, at which point things started to change rapidly. Apple changed the mobile phone business model with the creation of an open marketplace for smartphone software (the App Store) that was not controlled by the mobile operators. The appeal of the iPhone and the ease of constructing and installing new applications led to a growing demand for cost-effective mobile bandwidth as users wanted music, photos and then video. That in turn drove the 4G/LTE buildout that makes broadband smartphone applications broadly usable. Finally, at more or less the same time, Amazon introduced on-demand, pay-as-you go virtual computer resources and cloud platforms provided a means of having cloud backends for mobile applications.
So What, Now What?
So what’s the problem, you might well ask? Things work pretty well; we have high-bandwidth mobile devices with a plethora of interesting, internet-connected applications. The problem is the missed opportunity — the roughly trillion dollar cellular infrastructure that we’re not really taking advantage of except to provide raw Internet bandwidth. We’re not taking advantage of it because of the nature of the cellular infrastructure — closed, centrally managed, carefully optimized — and because of the very different business models behind cloud/ Internet and the mobile world.
MobiledgeX is addressing that opportunity head-on. We’re building a software layer that you can think of as running between the cloud and the mobile infrastructure that’s carefully designed and operated to be friendly and familiar to each side.
MobiledgeX is addressing that opportunity head-on. We’re building a software layer that you can think of as running between the cloud and the mobile infrastructure that’s carefully designed and operated to be friendly and familiar to each side. To an application developer, we look like flexible and easy-to-use cloud resources with some new and exciting features that come from the integration with the mobile operator networks. To mobile operators we’re a trusted partner – the code that we create to integrate with their infrastructure is carefully designed and tested according to mobile standards and practices. We present a familiar business model to the mobile operators if you think of how telephony roaming is done today. The mobile operators stand up extra capacity in both cases – telephony capacity beyond what their subscribers can use in the case of roaming; a little extra computational capacity in the case of MobiledgeX. We have a formal business relationship with each operator in our federation. When we use their resources they get paid. The resources could be access to unique compute locations and access to service endpoints for client developers, that execute as close to the end user as possible and data services, providing content relevant awareness.
I think there are lots of benefits. During the four years before I joined MobiledgeX, I led Ericsson’s cloud and virtualization efforts, and for nearly a decade before that I was the founder and CTO of Joyent, an early and innovative cloud services provider. I’ve been thinking about this for quite a while already.
We’ll talk in more detail about what MobiledgeX brings to the party, and how we enable them, in months to come, but for now let me mention a few.
For the simplest view of the opportunity, think about “edge” computing — how to provide computing and services nearer to the edge of the Internet. There are many anticipated uses of the edge; the hard thing is building it: where to you put those computational resources? Who owns and operates them?
The amazing opportunity for telcos is that, in their case, we know the answers. The global mobile infrastructure is run by a complex, distributed “control plane” that uses lots of standard X86 servers. All network service operators, wired and cellular, are moving from purpose-built appliances to software virtual forms of the functionality (that run on standard X86) servers. Our roadmap for providing powerful edge resources through our operator partnerships is right in front of us.
MobiledgeX also makes mobile infrastructure much more programmable and not so closed. That’s going to make a big difference in being to measure mobile Internet access performance and provide optimization services like CDN’s. Today performance analysis and optimization are sort of like on the wired Internet, twenty years ago.
Mobile operators are excited about the new programmability of their infrastructure because it lets them use these edge resources to augment cellular devices far beyond what can be done from today’s centralized cloud. Think of what that might mean in terms of valuable services that enable better augmented reality glasses or better location aware smartphone applications that know exactly where you’re pointing your phone camera at each instance.
And then finally we’re going to enable integration of applications with the local cellular infrastructure — a whole new set of applications to be invented. We can’t imagine all that might be done here and we look forward to enabling the developer community to find them. We’ve seen a few already by working with lead customers that need to use local cellular resources and edge computing. We are sure that is just the tip of the iceberg.
I’ve tried to keep this brief. There is a lot more to tell you about in times to come. Even more importantly, we think its developers and users that will discover most of the value of better cellular integration, not the mobile operators or companies like MobiledgeX. So if this triggers ideas and questions please contact us. We would love to hear from you.