This chapter presents a new vision of network operations, the self-driving network, that takes automation to the next level. This is not a description of existing work; rather, it is a challenge to dramatically rethink how we manage networks (or rather, how we do not manage networks). It draws upon an analogy with the development of self-driving cars and presents motivations for this effort. It then describes the technologies needed to implement this and an overall architecture of the system. As this endeavor will cause a major shift in network management, the chapter offers an evolutionary path to the end goal. Some of the consequences and human impacts of such a system are touched upon. The chapter concludes with some research topics and a final message. Key takeaways are that machine learning and feedback loops are fundamental to the solution; a key outcome is to build systems that are adaptive and predictive, for the benefit of users.
TopIntroduction
Advances in autonomous systems1 are all around us. As processors become more powerful and more energy efficient; as data becomes more ubiquitous and purposeful; and as software gets more sophisticated; our ability to hand over control to machines increases. In a very real sense, this is the logical conclusion of automation, the progression being: do things manually; find (sub-)tasks that are repetitive and automate them; and eventually, give the whole job over to a machine. This process requires considerable technological progress, as well as human considerations. Both of these will be explored in this chapter. A useful intermediate stage between automation and full autonomy is “augmentation,” where man and machine cooperatively operate a system; this will be discussed as well.
The very visible face of autonomous systems today is the self-driving car. Here again, we have gone from a very manual approach, where humans control every aspect of driving, to automating various driving functions; and from this, to creating fully autonomous vehicles. Efforts to automate driving functions have been sprinkled throughout the car’s 130-year history; efforts to build self-driving cars are much more recent. A significant trigger in the latter came from the Defense Advanced Research Projects Agency’s Grand Challenge to build an autonomous ground vehicle, held in 2004. Fifteen teams participated, albeit unsuccessfully; the following year, though, 5 teams were successful. This journey since has been long, but finally appears quite near technological success; now, human considerations dominate. The author will draw on this analogy in the discussion of self-driving networks; there are valuable lessons to learn in so doing, while bearing in mind that no analogy is perfect.