Abstract: Today's internet, accessed by fixed and mobile networks, allows us to transmit files, voice and video across the planet. With the emergence of an ultra-responsive and reliable ‘Tactile Internet,' advanced techniques in robotics and artificial intelligence, we predict the emergence of an ‘Internet of Skills' which allows the transmission of labor globally. It will invoke an important shift from content-delivery to skillset-delivery networks, where engineers would service cars or surgeons performing critical operations anywhere on the planet. For this to work, however, we require some fundamental laws of physics to be "reengineered." This keynote will discuss the vision, technology and building blocks for said emerging Internet of Skills. I will deep-dive into some of these building blocks, such as 5G and its ability to deliver ultra-low latency networking capabilities as well as open challenges from a compute and networking point of view.

Abstract: The term "distributed interactive application" today seems old-fashioned today since Cloudification has even turned humble word processing tasks into distributed interactive activities. Consequently, the property of being both interactive and distributed is frequently forgotten to the detriment of end users, who experience a lack of performance. Among gamers, the term "lag" has been coined for this, a catch-all phrase for all problems that end users experience as delay while interacting with an application.

The simplified term can hide excessive round-trip times, but it could also hide bandwidth limitations, processing delays or virtualisation overhead. In fact, any attempt at dealing with delay does first require that we understand the multitude of delays that can affect experience. Applications may be affected negatively by startup delays, rebuffering delays, rendering delays, processing delays, per-byte latency or load finishing time. The relevant kind of delay is application-specific, while the impact on user experience is application- as well as context-specific.

Research and development are continuously working to make improvements to all layers that make up distributed systems. Proposals for new generic architectures or architectural changes, can easily focus on a few of these faces of delay, satisfying a set of applications, while forgetting others that are equally valid .

The keynote attempts to bring structure to the many views on delay. We structure delay demands of applications by the kind of delay that matters to them and provide an insight into user experience studies for several of them. From there, we cast light onto recent and ongoing infrastructure developments, especially at the network and transport layer, and discuss their potential for fulfilling delay requirements.