A Large Scale Study of Data Center Network Reliability : the authors study reliability within and between Facebook datacenters. One of the key findings is the growth in complexity, heterogeneity and interconnectedness of datacenter increase the rate of occurrence of unwanted behaviours. Moreover, this seems to be also a key potential limiting factor for world scale spanning infrastructure undergoing rapid organic growth.
Failure Mode Effects Analysis (FMEA) : once you start reaching a certain production scale and more stringent requirement kicks in ( unless you were unlucky enough to have them at the get-go). You might want to run a failure modes and effects analysis (FMEA) is a step-by-step approach for identifying all possible failures in a design, a manufacturing or assembly process, or a product or service. While it was mainly designed to address shortcomings in the manufacturing industry, it is still extremely useful for IT system analysis, especially when you want to prepare yourself pre-rollout of a chaos monkey like system.
Applied AI hard problems : current and future AI hard problem, the interesting bit is the "emergent" behavior aspect that computer scientist are trying to achieve. Where AI is not tailored for a specific problem by adapt to the environment it encounter.
Dark silicon & Hardware Reliability : the authors look at the impact of the dark silicon approach ( when not all component are turned on when the system is up) and how to leverage the "dark" ratio to maximise lifespan of hardware. [slides]
TLDK : project lead by Intel within the fd.io framework. It is trying to adresse the lack of high level ( as in layer 4 ) packet processing capabilities. The project aim at delivering UDP/TCP etc.. packet processing on top of vector packet processing of FD.io (which can works on top of DPDK). By doing so Intel will be able to finally have a comprehensive framework which will enable DPDK based solution to flourish beyond the pure networking stack (NFV) solution.
Resilient ad serving at Twitter-scale : its interesting to see that their is a correlation between query latency and revenue for ad serving. This stem from the fact that latency for answering an ad query is dependent of the number of participant in the auction,and obviously the more participant the higher the revenue. However with the increased latency the higher the risk is to time out and hence revenue loss. Twitter use an adaptative system in order to maximise revenue while maintaining resiliency (availability), scalability, resource-utilization.
