1 / 10

Migrating Protocols In Multi-Threaded Message-Passing Systems

Migrating Protocols In Multi-Threaded Message-Passing Systems. Austin Anderson, HotSWUp ‘09. Safe Dynamic Software Update. Want preserve high level behaviors Message passing is one paradigm for inter-thread communication

nantai
Télécharger la présentation

Migrating Protocols In Multi-Threaded Message-Passing Systems

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Migrating Protocols In Multi-Threaded Message-Passing Systems Austin Anderson, HotSWUp ‘09

  2. Safe Dynamic Software Update • Want preserve high level behaviors • Message passing is one paradigm for inter-thread communication • We can guarantee that a system of parallel threads pass messages according to a specified protocol

  3. Session Types G ::= p →p : < T > Send a value of type T from sender p to recipient(s) p | G.G sequential comp | μt.G recursion | t recursive var

  4. Math Server Example G = p1→ p2 : < int > p1→ p2 : < int > p2→ p1 : < int> p1 p2 int send receive int send receive int receive send

  5. If we update p1 and p2 separately p1 p2 int send receive int send ERROR! no receive int receive send

  6. Update Coordination • Relies on static analysis which guarantees that after an update to a specific thread it will perform the new protocol • Work is completed in a submitted paper - copies available • To safely update the whole system we have to coordinate these individual updates

  7. Update Coordination Mechanism • Asynchronous global queue for communication • Messages sent under the old protocol will be consumed before messages sent under the new protocol become visible • We ensure that for every send that we perform a receive

  8. Producer Consumer Example p1 p2 G =μt.p1→ p2 : < int > p1→ p2 : < int > p1→ p2 : < bool > t μt. int receive send int receive send bool send receive t Queue 2 3 t 4 5 f 6 7 t 8 f

  9. Future Work • Preserving behaviors in multi-threaded shared state access • e.g. a piece of mutable state which can be read by multiple threads but only written by one (MRSW) • Generalised resource usage preservation which can represent • message passing • the above shared state which is accessed according to some pattern

  10. Conclusion • We can use a static analysis to guarantee each individual thread will have the new protocol after an update • System update safety is a coordination problem • Future work will generalize this to other high level behaviours/resource usages

More Related