query answers. 17 ??? Vasiliki Kalavri | Boston University 2020 Which tuples to drop? • Window-aware load shedding applies shedding to entire windows instead of individual tuples • When discarding • Naive approaches can lead to system instability or unnecessary load shedding. • In window-aware load shedding, queries need to define a batch size: an application-specific maximum tolerance to Stonebraker. Load shedding in a data stream manager. (VLDB ’03) • N. Tatbul and S. Zdonik. Window-aware load shedding for aggregation queries over data streams. (VLDB’06) • N. Tatbul, U. Çetintemel

RPC if the consumer exposes a service on the network • Failure handling: application needs to be aware of message loss, producers and consumers always online 5 Message queues • Asynchronous point-to-point

understanding of how stream processing systems work and what factors affect their performance • be aware of the challenges and trade-offs one needs to consider when designing and deploying streaming applications

Streaming state 3 • Explicit state primitives including state types and interfaces • The system is aware of state and can transparently checkpoint it, restore it, re- scale it Unmanaged Managed What

Pipelined Stream Filters. SIGMOD 2004. • Scheduling and placement • Peter R. Pietzuch et. al. Network-Aware Operator Placement for Stream-Processing Systems. ICDE 2006. • Brian Babcock et. al. Chain : Operator