I attended the Second IEEE WoWMoM Workshop on the Internet of Things: Smart Objects and Services, which took place last week, on June 4^th, in Madrid. The workshop opened with a keynote speech about SmartSantander, a Smart City project that successfully deployed a pilot consisting on different services in the city of Santander, Spain. The main takeaway of the presentation was the big challenge related to the integration of diverse data sources, ontologies and semantics. Such remarks set the trend of the workshop, where data management and aggregation were widely discussed. An interesting work was presented by Bob Ran from TNO, “Virtual Object Access Rights to Enable Multi-Party Use of Sensors” regarding the integration of different Sensors and Actuator Networks (SAN) and the possibility to buy data from already deployed networks instead of duplicating the sensing infrastructure.

The main purpose of the visit was to present our findings in “The M2M Promise, What Could Make it Happen? A Techno-economic Analysis” where we conclude that the proliferation of M2M solutions is not facing technical limitations; instead, solutions and services are largely hampered due to the uncertainty in the generation of value for major players involved in the study. Our message was well accepted, even though we target a highly technical audience, and interesting discussions where held during the remaining conference days.

The day after the workshops opened with a keynote by Mischa Dohler, where many of our findings were reinforced with his experience in trying to get M2M into de real world, not as pilot projects, but as actual services. Strong remarks fell again into the significance of data; how the obsession for privacy and security is only creating barriers in the information sharing, without the proper creation of intelligent data filters. During the presentation it was pointed out that before the huge information exchange takes place, efforts should be focused on M2M solutions that solves issues and gives revenues in the short term, which is exactly the same message that our study conveys.

The feeling after attending to the conference is that data science will only gain more relevance in the upcoming years and it will become a centerpiece of the Internet of things.

I also attended Globecom this year in Anaheim, CA, on Thu and Fri and mainly focused on following the recent advances in energy-efficient network designs.

On Thur morning, I presented a research paper about Energy-Efficient MU-MIMO in the “Energy Efficiency in Access Networks” session and listened to several other presentations.  The research results presented in this session covered issues from PHY layer all the way up to network deployment, which is expected as energy is consumed by protocols of all layers. While most of them are based on well-known ideas like dynamic/static BS management, power allocation, sleep mode, etc, there’re some interesting new concepts. For example, the sleep mode can be improved to consider the QoS (though I’m questioning if a device is sleeping, who will care about the QoS… ).  A student from Cioffi’s group presented a paper about intelligent power save mode design for devices connected to 802.11 WLAN. The battery life can be improved more than 1/3 especially when the packet arrival rate is low and the packet idle durations are long. It shows a good way to go when we consider connecting IoT devices through WLAN.

On the afternoon, I chaired a session, together with Oliver Blume from Bell Labs, on Green Cellular Wireless Communications. This session was even more interesting. There’re many new ideas about what we can do to save energy consumption for cellular networks. For example, energy-efficient beamforming, EE femtocell power optimization, utilization of stochastic geometry, and dynamic system bandwidth management. The interests came from both academia and industry and there were lots of questions/discussions in this sessions. The presentation by Oliver was especially interesting (slides attached below). He showed that by adapting the system bandwidth and # of operating circuits according to traffic load, a significant amount of energy can be saved.  The difficulties however lie in the design and implementation of adaptive hardware, which were unfortunately not discussed. This is one way to go, but definitely the only one. We should pursue more better solutions.

Globecom_2012_OliverBlume_clean

Friday was for tutorials and there were some good ones. For example, “Interference Alignment: State of the Art” by Syed A. Jafar and “Opportunistic Communication: Unified View and New Applications” by Aria Nosratinia, both cover the respective areas very well. However after some offline discussions, it seems the interference alignment doesn’t gain much momentum in industry because of the high complexity in the implementation. But will complexity always be a bottleneck? We have Moore’s law…

On Fri afternoon, I presented a tutorial about Joint PHY-MAC Design for Spectral- and Energy-Efficient Wireless Networks. This tutorial summarizes  spectral and energy efficient communication technologies for both individual  and multi-user networks.