Wireless data networks: issues beyond the link layer

Summary form only given. The emergence of the WWW and the underlying Internet forever changed the way in which hundreds of millions of people access information and conduct online transactions. However, for the large part, this access requires fairly traditional computers and wired network access. Wireless network access and non-traditional handheld computing devices have the potential to increase both the reach and the impact of the WWW, and the online information access model, by an order of magnitude or more. Together, wireless networks and mobile computing will help in realizing the dream of truly ubiquitous access to the information superhighway. Of course, there are many technical problems standing in the way of this vision, and this paper will discuss some of these problems. While many of the difficult problems are indeed unique to the lower layers of the wireless link, such as modulation, MAC protocols, etc., we will focus on a separate set of issues. The problems we will focus on start from the well known one of extending IP routing to wireless hosts, and then go on to look at congestion control protocols such as TCP. We then will examine the emerging area of service discovery protocols and finally go on to describe a content adaptation framework that acts as an impedance matcher between the high-speed wired data world and the lower-speed wireless world. The questions that need to be asked in all these domains is whether standard Internet protocols apply well to the wireless data world, do they need to be extended or should they be replaced? Different researchers have taken different positions in the past, and we will explore when these different positions make sense.