Building the Right Mobile Broadband Infrastructure for the Cloud Age

The growth of mobile broadband is putting pressure on network capacity and profitability. Operators are seeking the best way to move forward, and smart pipe is one of these ways. ZTE convened a forum on smart pipes on February 29, the third day of MWC 2012. Bill Huang, general manager of the China Mobile Research Institute, gave a speech. He talked about what it takes to build an infrastructure that caters to the growth of mobile broadband.

It is my honor and pleasure to be here for the third time. I’d like to take this opportunity to congratulate ZTE for your success in the business and your contribution to promoting and developing TD-LTE.

I want to use the word “cloud age” in this speech because I truly believe the emergence of cloud computing is a unique opportunity for the IT industry to take the application platform away from the desktop PC and put it into portable mobile devices. This could be a true revolution not only for mobile communications but also for information technology in general.

There is one point I want to make: With the explosive growth of mobile broadband, it is very important for operators to recognize that new business will be centered on providing bandwidth. In many emerging economies, the lack of well-built 3G infrastructure and more expensive smartphones means bandwidth revenue is still reasonably small. If you look at vendors like Softbank, Hutchison, and NTT Docomo, the bulk of their revenue comes from data traffic. It’s very important, therefore, to transform our businesses and put value on traffic. Traffic is no longer going to be a supplementary service; it’s going to be our new business. When telephones are provided free of charge, this will be our source of income.

Another point I would make is this: If you look at some of the simple numbers China Mobile has collected from its network, there is a dramatic increase in data traffic. Data traffic easily doubles every year. It’s not hard to imagine that it will increase a thousand times in ten years. If we increase the traffic a thousand times, can we increase the ARPU a thousand times? The answer is obviously no. Can we increase the ARPU ten times? Perhaps no. Can we triple ARPU? Maybe. So the pressure for us to find ways of building the right infrastructure to support this business model is unprecedented. 

LTE is emerging as the main 4G technology. Specifically, I want to talk about TD-LTE. Throughout the world, TDD spectrums are the only spectrums that are still abundant, and the cost of these spectrums is considerably lower than FDD spectrums. The unique advantage of having asymmetric downlink and uplink makes TDD more economical; you get more bandwidth out of the same spectrum. China Mobile is currently deploying TDD on a very large scale. The entire TD-LTE ecosystem has already been built up and is running. Converged TDD-LTE and FDD-LTE is going to be the most ubiquitous RF standard in history. At the global TD-LTE initiative (GTI) conference, held two days ago, our CEO predicted that China Mobile is going to deploy 20,000 base stations by the end of 2012 and 200,000 base stations globally by 2014. That is a fundamental shift towards LTE.  

For our bandwidth business, we can’t ignore cost. Nor can we ignore ubiquity of terminals. Wi-Fi or WLAN is now emerging as an essential tool for mobile operators. We have already included WLAN as part of our mobile networks for China Mobile, not only as a supplementary thing. If you look at the numbers, we have already deployed close to 1.6 million access points. In three years, six million APs will be deployed. Where will we deploy these APs? We will place them in hot zones or hotspots where people are most likely to use mobile broadband in stationary conditions. We will eventually place them in residential areas, schools, and enterprises.  The natural evolution of this is international roaming, international tariffs, and international authentication for WLAN. WLAN worldwide roaming is a big issue. People don’t know how to select the right Wi-Fi points, and payment is also complicated.

What would be the backbone for this new network? We hope to achieve 10 Mbps for personal users, 100 Mbps for family users, and 1000 Mbps for enterprises. To accomplish this, we must quickly migrate SDH-based transport networks to packet-based transport networks. We call this SDH-PTN migration. The new hierarchy becomes 1G-10G-100G. 40G is being used right now, but we have seen the economies of the target countries for 100G strengthening, and it is becoming more feasible to push 100G into commercialization. At first, 100G may be more expensive to deploy than 40G, but as 100G becomes more popular, it becomes more economical because fiber is saved.

What is the true goal of building smart pipes? Building the smart pipes means creating the mechanism to extract value from bandwidth businesses. The dumbest pipe is the Internet, where all applications have equal priorities. P2P movie download is the same as Skype, and this doesn’t make sense. If we follow this business model, we will go out of business. The appetite for bandwidth is insatiable; if we double our international gateway from 100 Gbps to 200 Gbps, P2P software, which is constantly testing bandwidth availability, will very quickly discover this, and your pipe will soon be full. Without the proper smart pipe technology for policy awareness, traffic shaping, caching, and streaming, the network will always be congested.

Network congestion is inherent in the current design of the Internet. It is very important to the future of the IP network and Internet that we use policy management to create SLAs. We potentially need to apply policy control based on SLA in smart pipes. Today it is missing. We have to understand in the mobile broadband age that always-online is increasingly important, but it is an increasing headache for mobile network design. 

Signaling has been and continues to be an uncharged overhead when users access the network. If signaling dominates network traffic, the user will only be able to transmit one byte. Think about the nightmare this will cause. You build the network and create the signaling so you can charge people to transmit at 1 Mbps, but instead, people can only send one byte and you use maybe 100 bytes or 1000 bytes to help that one byte. That’s why the network is going crazy. This is a very important issue to be resolved. Increasingly, operators are building always-online facilities, that is, building software and API to enable applications so that their heartbeat becomes organized and not random.

New technology is also creating problems for our network construction. One very important new technology I want to talk about is cloud-based radio access network (C-RAN). In C-RAN, baseband processing is centralized. The radio and antennas, through which we transmit digitized signals over fiber, are distributed. This is a big issue because the speed of the backhaul is so high that most people are using naked fiber. Can we do this on a large scale? The answer is no. I don’t know how we can solve this problem, but we need to work on it. If cloud RAN is to be successful, we need to solve the backhaul for cloud RAN signals. The challenge is there. There are potentially many different ways to solve this problem. You can compress the radio signal so that you have less bandwidth, or you can make cheaper optics. In any case, this is the future if we want to save 15% CAPEX and 18% OPEX. To achieve this, we have to do something.

I think the future is bright for the mobile industry. It is not only bright but burning. We have to be ready for that high growth. Now we want to increase usage, but we want to increase revenue at the same time. I hope we can work together with operators, developers, and vendors. We are focusing our energy on solving these problems and creating better business models.