coordination free computations
play

Coordination-Free Computations Christopher Meiklejohn LASP - PowerPoint PPT Presentation

Sep 08, 2022 •27 likes •1.37k views

Coordination-Free Computations Christopher Meiklejohn LASP DISTRIBUTED, EVENTUALLY CONSISTENT COMPUTATIONS CHRISTOPHER MEIKLEJOHN (BASHO TECHNOLOGIES, INC.) PETER VAN ROY (UNIVERSIT CATHOLIQUE DE LOUVAIN) 2 LASP MOTIVATION 3

  1. Increment Rovio Ad Counter1 Counter2 Counter3 Counter Read Remove Read ≥ 50,000 Rovio Ads Rovio Ad Counter1 Counter2 Counter3 Counter Ads Ads Union Product Filter Ads With Contracts Contracts Riot Ad Counter1 Counter2 Counter3 Counter Lasp Operation Contracts Riot Ads User-Maintained CRDT Riot Ad Counter1 Counter2 Counter3 Counter Lasp-Maintained CRDT 67

  2. Rovio Ad Counter Union Product Ads Riot Ad Counter Contracts Riot Ads Riot Ad Counter 68

  3. Increment Rovio Ad Counter Read Remove ≥ 50,000 Rovio Ads Rovio Ad Counter Ads Union Product Filter Ads Contracts Riot Ad Counter Lasp Operation Contracts Riot Ads User-Maintained CRDT Riot Ad Counter 69

  4. Ads Ads Union Product Filter Ads With Contracts Contracts Lasp Operation Contracts User-Maintained CRDT 70

  5. Read Ads Ads Product Filter Ads With Contracts Contracts Lasp Operation Contracts 71

  6. Counter1 Counter2 Counter3 Read Counter1 Counter2 Counter3 Ads Ads Filter With Contracts Contracts Counter1 Counter2 Counter3 Lasp Operation User-Maintained CRDT Counter1 Counter2 Counter3 Lasp-Maintained CRDT 72

  7. Increment Rovio Ad Counter1 Counter2 Counter3 Counter Read Remove Read ≥ 50,000 Rovio Ads Rovio Ad Counter1 Counter2 Counter3 Counter Ads Ads Union Product Filter Ads With Contracts Contracts Riot Ad Counter1 Counter2 Counter3 Counter Lasp Operation Contracts Riot Ads User-Maintained CRDT Riot Ad Counter1 Counter2 Counter3 Counter Lasp-Maintained CRDT 73

  8. Increment Rovio Ad Counter Read Remove ≥ 50,000 Rovio Ads Rovio Ad Counter Ads Union Product Ads Contracts Riot Ad Counter Lasp Operation Contracts Riot Ads 74

  9. Counter1 Counter2 Counter3 Read Remove Read 50,000 Rovio Ads Counter1 Counter2 Counter3 Ads Ads Union Product Filter Ads With Contracts Contracts Counter1 Counter2 Counter3 Lasp Operation Contracts Riot Ads User-Maintained CRDT Counter1 Counter2 Counter3 Lasp-Maintained CRDT 75

  10. INFORMATION FLOW: MONOTONIC METADATA TO PREVENT DUPLICATE PROPAGATION 76

  11. AD COUNTER EXAMPLE CODE 77

  12. %% @doc Client process; standard recursive looping server. client(Id, AdsWithContracts, PreviousValue) -> receive view_ad -> %% Get current ad list. {ok, {_, _, AdList0}} = lasp:read(AdsWithContracts, PreviousValue), AdList = riak_dt_orset:value(AdList0), case length(AdList) of 0 -> %% No advertisements left to display; ignore %% message. client(Id, AdsWithContracts, AdList0); _ -> %% Select a random advertisement from the list of %% active advertisements. {#ad{counter=Ad}, _} = lists:nth( random:uniform(length(AdList)), AdList), %% Increment it. {ok, _} = lasp:update(Ad, increment, Id), lager:info("Incremented ad counter: ~p", [Ad]), client(Id, AdsWithContracts, AdList0) end end.

  13. %% @doc Server functions for the advertisement counter. After 5 views, %% disable the advertisement. %% server({#ad{counter=Counter}=Ad, _}, Ads) -> %% Blocking threshold read for 5 advertisement impressions. {ok, _} = lasp:read(Counter, 5), %% Remove the advertisement. {ok, _} = lasp:update(Ads, {remove, Ad}, Ad), lager:info("Removing ad: ~p", [Ad]).

  14. %% Generate a series of unique identifiers. %% Compute the Cartesian product of both ads and contracts. RovioAdIds = lists:map(fun(_) -> druuid:v4() end, lists:seq(1, 10)), {ok, AdsContracts} = lasp:declare(?SET), lager:info("Rovio Ad Identifiers are: ~p", [RovioAdIds]), ok = lasp:product(Ads, Contracts, AdsContracts), TriforkAdIds = lists:map(fun(_) -> druuid:v4() end, lists:seq(1, 10)), %% Filter items by join on item it. lager:info("Trifork Ad Identifiers are: ~p", [TriforkAdIds]), {ok, AdsWithContracts} = lasp:declare(?SET), FilterFun = fun({#ad{id=Id1}, #contract{id=Id2}}) -> Ids = RovioAdIds ++ TriforkAdIds, Id1 =:= Id2 lager:info("Ad Identifiers are: ~p", [Ids]), end, ok = lasp:filter(AdsContracts, FilterFun, AdsWithContracts), %% Generate Rovio's advertisements. {ok, RovioAds} = lasp:declare(?SET), %% Launch a series of client processes, each of which is responsible lists:map(fun(Id) -> %% for displaying a particular advertisement. %% Generate a G-Counter. {ok, CounterId} = lasp:declare(?COUNTER), %% Generate a OR-set for tracking clients. %% Add it to the advertisement set. {ok, Clients} = lasp:declare(?SET), {ok, _} = lasp:update(RovioAds, {add, #ad{id=Id, counter=CounterId}}, %% Each client takes the full list of ads when it starts, and reads undefined) %% from the variable store. lists:map(fun(Id) -> end, RovioAdIds), ClientPid = spawn_link(?MODULE, client, [Id, AdsWithContracts, undefined]), %% Generate Trifork's advertisements. {ok, _} = lasp:update(Clients, {ok, TriforkAds} = lasp:declare(?SET), {add, ClientPid}, lists:map(fun(Id) -> undefined) %% Generate a G-Counter. end, lists:seq(1,5)), {ok, CounterId} = lasp:declare(?COUNTER), %% Add it to the advertisement set. %% Launch a server process for each advertisement, which will block {ok, _} = lasp:update(TriforkAds, %% until the advertisement should be disabled. {add, #ad{id=Id, counter=CounterId}}, undefined) %% Create a OR-set for the server list. {ok, Servers} = lasp:declare(?SET), end, TriforkAdIds), %% Get the current advertisement list. %% Union ads. {ok, {_, _, AdList0}} = lasp:read(AdsWithContracts), {ok, Ads} = lasp:declare(?SET), AdList = riak_dt_orset:value(AdList0), ok = lasp:union(RovioAds, TriforkAds, Ads), %% For each advertisement, launch one server for tracking it's %% For each identifier, generate a contract. %% impressions and wait to disable. {ok, Contracts} = lasp:declare(?SET), lists:map(fun(Ad) -> lists:map(fun(Id) -> ServerPid = spawn_link(?MODULE, server, [Ad, Ads]), {ok, _} = lasp:update(Contracts, {ok, _} = lasp:update(Servers, {add, #contract{id=Id}}, {add, ServerPid}, undefined) undefined) end, Ids), end, AdList),

  15. %% Generate a series of unique identifiers. %% Compute the Cartesian product of both ads and contracts. RovioAdIds = lists:map(fun(_) -> druuid:v4() end, lists:seq(1, 10)), {ok, AdsContracts} = lasp:declare(?SET), lager:info("Rovio Ad Identifiers are: ~p", [RovioAdIds]), ok = lasp:product(Ads, Contracts, AdsContracts), TriforkAdIds = lists:map(fun(_) -> druuid:v4() end, lists:seq(1, 10)), %% Filter items by join on item it. lager:info("Trifork Ad Identifiers are: ~p", [TriforkAdIds]), {ok, AdsWithContracts} = lasp:declare(?SET), FilterFun = fun({#ad{id=Id1}, #contract{id=Id2}}) -> Ids = RovioAdIds ++ TriforkAdIds, Id1 =:= Id2 lager:info("Ad Identifiers are: ~p", [Ids]), end, ok = lasp:filter(AdsContracts, FilterFun, AdsWithContracts), %% Generate Rovio's advertisements. {ok, RovioAds} = lasp:declare(?SET), %% Launch a series of client processes, each of which is responsible lists:map(fun(Id) -> %% for displaying a particular advertisement. %% Generate a G-Counter. {ok, CounterId} = lasp:declare(?COUNTER), %% Generate a OR-set for tracking clients. %% Add it to the advertisement set. {ok, Clients} = lasp:declare(?SET), {ok, _} = lasp:update(RovioAds, {add, #ad{id=Id, counter=CounterId}}, %% Each client takes the full list of ads when it starts, and reads undefined) %% from the variable store. lists:map(fun(Id) -> end, RovioAdIds), ClientPid = spawn_link(?MODULE, client, [Id, AdsWithContracts, undefined]), %% Generate Trifork's advertisements. {ok, _} = lasp:update(Clients, {ok, TriforkAds} = lasp:declare(?SET), {add, ClientPid}, lists:map(fun(Id) -> undefined) %% Generate a G-Counter. end, lists:seq(1,5)), {ok, CounterId} = lasp:declare(?COUNTER), %% Add it to the advertisement set. %% Launch a server process for each advertisement, which will block {ok, _} = lasp:update(TriforkAds, %% until the advertisement should be disabled. {add, #ad{id=Id, counter=CounterId}}, undefined) %% Create a OR-set for the server list. {ok, Servers} = lasp:declare(?SET), end, TriforkAdIds), %% Get the current advertisement list. %% Union ads. {ok, {_, _, AdList0}} = lasp:read(AdsWithContracts), {ok, Ads} = lasp:declare(?SET), AdList = riak_dt_orset:value(AdList0), ok = lasp:union(RovioAds, TriforkAds, Ads), %% For each advertisement, launch one server for tracking it's %% For each identifier, generate a contract. %% impressions and wait to disable. {ok, Contracts} = lasp:declare(?SET), lists:map(fun(Ad) -> lists:map(fun(Id) -> ServerPid = spawn_link(?MODULE, server, [Ad, Ads]), {ok, _} = lasp:update(Contracts, {ok, _} = lasp:update(Servers, {add, #contract{id=Id}}, {add, ServerPid}, undefined) undefined) end, Ids), end, AdList),

  16. %% Generate a series of unique identifiers. %% Compute the Cartesian product of both ads and contracts. RovioAdIds = lists:map(fun(_) -> druuid:v4() end, lists:seq(1, 10)), {ok, AdsContracts} = lasp:declare(?SET), lager:info("Rovio Ad Identifiers are: ~p", [RovioAdIds]), ok = lasp:product(Ads, Contracts, AdsContracts), TriforkAdIds = lists:map(fun(_) -> druuid:v4() end, lists:seq(1, 10)), %% Filter items by join on item it. lager:info("Trifork Ad Identifiers are: ~p", [TriforkAdIds]), {ok, AdsWithContracts} = lasp:declare(?SET), FilterFun = fun({#ad{id=Id1}, #contract{id=Id2}}) -> Ids = RovioAdIds ++ TriforkAdIds, Id1 =:= Id2 lager:info("Ad Identifiers are: ~p", [Ids]), end, ok = lasp:filter(AdsContracts, FilterFun, AdsWithContracts), %% Generate Rovio's advertisements. {ok, RovioAds} = lasp:declare(?SET), %% Launch a series of client processes, each of which is responsible lists:map(fun(Id) -> %% for displaying a particular advertisement. %% Generate a G-Counter. {ok, CounterId} = lasp:declare(?COUNTER), %% Generate a OR-set for tracking clients. %% Add it to the advertisement set. {ok, Clients} = lasp:declare(?SET), {ok, _} = lasp:update(RovioAds, {add, #ad{id=Id, counter=CounterId}}, %% Each client takes the full list of ads when it starts, and reads undefined) %% from the variable store. lists:map(fun(Id) -> end, RovioAdIds), ClientPid = spawn_link(?MODULE, client, [Id, AdsWithContracts, undefined]), %% Generate Trifork's advertisements. {ok, _} = lasp:update(Clients, {ok, TriforkAds} = lasp:declare(?SET), {add, ClientPid}, lists:map(fun(Id) -> undefined) %% Generate a G-Counter. end, lists:seq(1,5)), {ok, CounterId} = lasp:declare(?COUNTER), %% Add it to the advertisement set. %% Launch a server process for each advertisement, which will block {ok, _} = lasp:update(TriforkAds, %% until the advertisement should be disabled. {add, #ad{id=Id, counter=CounterId}}, undefined) %% Create a OR-set for the server list. {ok, Servers} = lasp:declare(?SET), end, TriforkAdIds), %% Get the current advertisement list. %% Union ads. {ok, {_, _, AdList0}} = lasp:read(AdsWithContracts), {ok, Ads} = lasp:declare(?SET), AdList = riak_dt_orset:value(AdList0), ok = lasp:union(RovioAds, TriforkAds, Ads), %% For each advertisement, launch one server for tracking it's %% For each identifier, generate a contract. %% impressions and wait to disable. {ok, Contracts} = lasp:declare(?SET), lists:map(fun(Ad) -> lists:map(fun(Id) -> ServerPid = spawn_link(?MODULE, server, [Ad, Ads]), {ok, _} = lasp:update(Contracts, {ok, _} = lasp:update(Servers, {add, #contract{id=Id}}, {add, ServerPid}, undefined) undefined) end, Ids), end, AdList),

  17. %% Generate a series of unique identifiers. %% Compute the Cartesian product of both ads and contracts. RovioAdIds = lists:map(fun(_) -> druuid:v4() end, lists:seq(1, 10)), {ok, AdsContracts} = lasp:declare(?SET), lager:info("Rovio Ad Identifiers are: ~p", [RovioAdIds]), ok = lasp:product(Ads, Contracts, AdsContracts), TriforkAdIds = lists:map(fun(_) -> druuid:v4() end, lists:seq(1, 10)), %% Filter items by join on item it. lager:info("Trifork Ad Identifiers are: ~p", [TriforkAdIds]), {ok, AdsWithContracts} = lasp:declare(?SET), FilterFun = fun({#ad{id=Id1}, #contract{id=Id2}}) -> Ids = RovioAdIds ++ TriforkAdIds, Id1 =:= Id2 lager:info("Ad Identifiers are: ~p", [Ids]), end, ok = lasp:filter(AdsContracts, FilterFun, AdsWithContracts), %% Generate Rovio's advertisements. {ok, RovioAds} = lasp:declare(?SET), %% Launch a series of client processes, each of which is responsible lists:map(fun(Id) -> %% for displaying a particular advertisement. %% Generate a G-Counter. {ok, CounterId} = lasp:declare(?COUNTER), %% Generate a OR-set for tracking clients. %% Add it to the advertisement set. {ok, Clients} = lasp:declare(?SET), {ok, _} = lasp:update(RovioAds, {add, #ad{id=Id, counter=CounterId}}, %% Each client takes the full list of ads when it starts, and reads undefined) %% from the variable store. lists:map(fun(Id) -> end, RovioAdIds), ClientPid = spawn_link(?MODULE, client, [Id, AdsWithContracts, undefined]), %% Generate Trifork's advertisements. {ok, _} = lasp:update(Clients, {ok, TriforkAds} = lasp:declare(?SET), {add, ClientPid}, lists:map(fun(Id) -> undefined) %% Generate a G-Counter. end, lists:seq(1,5)), {ok, CounterId} = lasp:declare(?COUNTER), %% Add it to the advertisement set. %% Launch a server process for each advertisement, which will block {ok, _} = lasp:update(TriforkAds, %% until the advertisement should be disabled. {add, #ad{id=Id, counter=CounterId}}, undefined) %% Create a OR-set for the server list. {ok, Servers} = lasp:declare(?SET), end, TriforkAdIds), %% Get the current advertisement list. %% Union ads. {ok, {_, _, AdList0}} = lasp:read(AdsWithContracts), {ok, Ads} = lasp:declare(?SET), AdList = riak_dt_orset:value(AdList0), ok = lasp:union(RovioAds, TriforkAds, Ads), %% For each advertisement, launch one server for tracking it's %% For each identifier, generate a contract. %% impressions and wait to disable. {ok, Contracts} = lasp:declare(?SET), lists:map(fun(Ad) -> lists:map(fun(Id) -> ServerPid = spawn_link(?MODULE, server, [Ad, Ads]), {ok, _} = lasp:update(Contracts, {ok, _} = lasp:update(Servers, {add, #contract{id=Id}}, {add, ServerPid}, undefined) undefined) end, Ids), end, AdList),

  18. %% Generate a series of unique identifiers. %% Compute the Cartesian product of both ads and contracts. RovioAdIds = lists:map(fun(_) -> druuid:v4() end, lists:seq(1, 10)), {ok, AdsContracts} = lasp:declare(?SET), lager:info("Rovio Ad Identifiers are: ~p", [RovioAdIds]), ok = lasp:product(Ads, Contracts, AdsContracts), TriforkAdIds = lists:map(fun(_) -> druuid:v4() end, lists:seq(1, 10)), %% Filter items by join on item it. lager:info("Trifork Ad Identifiers are: ~p", [TriforkAdIds]), {ok, AdsWithContracts} = lasp:declare(?SET), FilterFun = fun({#ad{id=Id1}, #contract{id=Id2}}) -> Ids = RovioAdIds ++ TriforkAdIds, Id1 =:= Id2 lager:info("Ad Identifiers are: ~p", [Ids]), end, ok = lasp:filter(AdsContracts, FilterFun, AdsWithContracts), %% Generate Rovio's advertisements. {ok, RovioAds} = lasp:declare(?SET), %% Launch a series of client processes, each of which is responsible lists:map(fun(Id) -> %% for displaying a particular advertisement. %% Generate a G-Counter. {ok, CounterId} = lasp:declare(?COUNTER), %% Generate a OR-set for tracking clients. %% Add it to the advertisement set. {ok, Clients} = lasp:declare(?SET), {ok, _} = lasp:update(RovioAds, {add, #ad{id=Id, counter=CounterId}}, %% Each client takes the full list of ads when it starts, and reads undefined) %% from the variable store. lists:map(fun(Id) -> end, RovioAdIds), ClientPid = spawn_link(?MODULE, client, [Id, AdsWithContracts, undefined]), %% Generate Trifork's advertisements. {ok, _} = lasp:update(Clients, {ok, TriforkAds} = lasp:declare(?SET), {add, ClientPid}, lists:map(fun(Id) -> undefined) %% Generate a G-Counter. end, lists:seq(1,5)), {ok, CounterId} = lasp:declare(?COUNTER), %% Add it to the advertisement set. %% Launch a server process for each advertisement, which will block {ok, _} = lasp:update(TriforkAds, %% until the advertisement should be disabled. {add, #ad{id=Id, counter=CounterId}}, undefined) %% Create a OR-set for the server list. {ok, Servers} = lasp:declare(?SET), end, TriforkAdIds), %% Get the current advertisement list. %% Union ads. {ok, {_, _, AdList0}} = lasp:read(AdsWithContracts), {ok, Ads} = lasp:declare(?SET), AdList = riak_dt_orset:value(AdList0), ok = lasp:union(RovioAds, TriforkAds, Ads), %% For each advertisement, launch one server for tracking it's %% For each identifier, generate a contract. %% impressions and wait to disable. {ok, Contracts} = lasp:declare(?SET), lists:map(fun(Ad) -> lists:map(fun(Id) -> ServerPid = spawn_link(?MODULE, server, [Ad, Ads]), {ok, _} = lasp:update(Contracts, {ok, _} = lasp:update(Servers, {add, #contract{id=Id}}, {add, ServerPid}, undefined) undefined) end, Ids), end, AdList),

  19. AD COUNTER DISTRIBUTION 85

  20. Increment Rovio Ad Counter1 Counter2 Counter3 Counter Read Remove Read ≥ 50,000 Rovio Ads Rovio Ad Counter1 Counter2 Counter3 Counter Ads Ads Union Product Filter Ads With Contracts Contracts Riot Ad Counter1 Counter2 Counter3 Counter Lasp Operation Contracts Riot Ads User-Maintained CRDT Riot Ad Counter1 Counter2 Counter3 Counter Lasp-Maintained CRDT 86

  21. Increment Rovio Ad Counter1 Counter2 Counter3 Counter Read Remove Read ≥ 50,000 Rovio Ads Rovio Ad Counter1 Counter2 Counter3 Counter Ads Ads Union Product Filter Ads With Contracts Contracts Riot Ad Counter1 Counter2 Counter3 Counter Lasp Operation Contracts Riot Ads User-Maintained CRDT Riot Ad Counter1 Counter2 Counter3 Counter Lasp-Maintained CRDT 87

  22. Increment Rovio Ad Counter1 Counter2 Counter3 Counter Read Remove Read ≥ 50,000 Rovio Ads Rovio Ad Counter1 Counter2 Counter3 Counter Ads Ads Union Product Filter Ads With Contracts Contracts Riot Ad Counter1 Counter2 Counter3 Counter Lasp Operation Contracts Riot Ads User-Maintained CRDT Riot Ad Counter1 Counter2 Counter3 Counter Lasp-Maintained CRDT 88

  23. Increment Rovio Ad Counter1 Counter2 Counter3 Counter Read Remove Read ≥ 50,000 Rovio Ads Rovio Ad Counter1 Counter2 Counter3 Counter Ads Ads Union Product Filter Ads With Contracts Contracts Riot Ad Counter1 Counter2 Counter3 Counter Lasp Operation Contracts Riot Ads User-Maintained CRDT Riot Ad Counter1 Counter2 Counter3 Counter Lasp-Maintained CRDT 89

  24. DISTRIBUTION BOUNDARIES: ARBITRARY ALLOWS COMPOSITION OF ENTIRE SYSTEM 90

  25. MATERIALIZED VIEWS 91

  26. MATERIALIZED VIEWS: INCREMENTALLY UPDATING, DELTA PROPAGATION SCHEMA BASED OR DYNAMIC; 2I IN RIAK 92

  27. DATABASE AS STREAM OF UPDATES 93

  28. K1 K2 K3 K1 K2 Riak 94

  29. RS1 K1 K2 K3 K1 K2 RS2 Riak RS3 95

  30. K1 K1 RS1 K1 K2 K3 K1 K2 K2 K2 RS2 Riak K3 RS3 96

  31. K1 K1 N1 N3 N2 K1 K2 K3 K1 K2 K2 K2 RS2 Riak K3 RS3 97

  32. COMPOSE PROGRAMS 98

  33. K1 K1 N1 N3 N2 K1 K2 K3 K1 K2 K2 K2 RS2 Riak K3 RS3 Lasp Stream Processor Template 99

  34. K1 K1 N1 N3 N2 K1 K2 K3 K1 K2 K2 K2 RS2 Riak K3 RS3 Lasp All Objects Program Lasp Stream Processor Template 100

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

NetChain: Scale-Free Sub-RTT Coordination Xin Jin Xiaozhou Li, Haoyu Zhang, Robert Soul,

NetChain: Scale-Free Sub-RTT Coordination Xin Jin Xiaozhou Li, Haoyu Zhang, Robert Soul,

NetChain: Scale-Free Sub-RTT Coordination Xin Jin Xiaozhou Li, Haoyu Zhang, Robert Soul, Jeongkeun Lee, Nate Foster, Changhoon Kim, Ion Stoica Conventional wisdom: avoid coordination NetChain: lightning fast coordination enabled by

484 views • 37 slides
Coordination-free query evaluation and multi-query optimization in parallel and distributed

Coordination-free query evaluation and multi-query optimization in parallel and distributed

Formal approaches to: Coordination-free query evaluation and multi-query optimization in parallel and distributed systems Bas Ketsman Outline CALM Formalization CALM Revision 1 Coordination-free evaluation Conclusion Parallel-Correctness

1.25k views • 88 slides
Embarrassingly Parallel Computations 3.2 1 Embarrassingly Parallel Computations A computation

Embarrassingly Parallel Computations 3.2 1 Embarrassingly Parallel Computations A computation

Parallel Techniques Embarrassingly Parallel Computations Partitioning and Divide-and-Conquer Strategies Pipelined Computations Synchronous Computations Asynchronous Computations Asynchronous Computations Strategies

378 views • 13 slides
ZooKeeper: Wait-free coordination for Internet-scale systems Xuyang Zhang Zhesheng Xie What is

ZooKeeper: Wait-free coordination for Internet-scale systems Xuyang Zhang Zhesheng Xie What is

ZooKeeper: Wait-free coordination for Internet-scale systems Xuyang Zhang Zhesheng Xie What is ZooKeeper A distributed coordination service for distributed applications that providing a simple and high performance kernel for building more

693 views • 37 slides
Asymptotic Improvement of Computations over Free Monads Janis Voigtl ander Technische

Asymptotic Improvement of Computations over Free Monads Janis Voigtl ander Technische

Asymptotic Improvement of Computations over Free Monads Janis Voigtl ander Technische Universit at Dresden MPC08 Monads for IO in Haskell Program: echo :: IO () echo = do c getChar when ( c = ) $ do putChar c echo

1.31k views • 90 slides
Coordination models Essence We are trying to separate computation from coordination; coordination

Coordination models Essence We are trying to separate computation from coordination; coordination

Coordination-Based Systems Coordination Models Coordination models Essence We are trying to separate computation from coordination; coordination deals with all aspects of communication between processes, as well as their cooperation.

453 views • 31 slides
Structuring Computations Structuring Computations Contents Jacobs Types06, 18/4/06

Structuring Computations Structuring Computations Contents Jacobs Types06, 18/4/06

FACULTY OF SCIENCE Bart Jacobs Structuring Computations Structuring Computations Contents Jacobs Types06, 18/4/06 p.1/52 Structuring Computations Contents I. Sneak preview VII. Hoare logic for JML II. Comonads VIII. Conclusions

1.79k views • 164 slides
Project Coordination of the social security systems in the frame of free movement of

Project Coordination of the social security systems in the frame of free movement of

Project Coordination of the social security systems in the frame of free movement of workers Short Presentation of Partners from the Social security schemes for the European Conference about developing the cooperation between agricultural

278 views • 15 slides
But why is Synchronous Java was designed THREADS concurrency coordination in the

But why is Synchronous Java was designed THREADS concurrency coordination in the

Modeling Concurrency with Actors A Journey into Erlang Land Eugne Delacroix, Hamlet and Horatio in the Graveyard Functional and Interactive concurrency Coordination is the new imperative The Free Lunch is Over What I will Tell You Develop

330 views • 5 slides
Zeno-free, distributed event-triggered coordination for multi-agent average consensus Cameron

Zeno-free, distributed event-triggered coordination for multi-agent average consensus Cameron

Zeno-free, distributed event-triggered coordination for multi-agent average consensus Cameron Nowzari 1 es 2 Jorge Cort 1 Electrical and Systems Engineering University of Pennsylvania cnowzari@seas.upenn.edu 2 Mechanical and Aerospace

688 views • 55 slides
The computations of acting agents and the agents acting in computations Philipp Hennig ICERM 5

The computations of acting agents and the agents acting in computations Philipp Hennig ICERM 5

The computations of acting agents and the agents acting in computations Philipp Hennig ICERM 5 June 2017 Research Group for Probabilistic Numerics Max Planck Institute for Intelligent Systems Tbingen, Germany Some of the presented work was

1.12k views • 76 slides
Embarrassingly Parallel Computations Embarrassingly Parallel Computations A computation that

Embarrassingly Parallel Computations Embarrassingly Parallel Computations A computation that

Embarrassingly Parallel Computations Embarrassingly Parallel Computations A computation that can be divided into completely independent parts, each of which can be executed on a separate process(or) is called embarrassingly parallel. An

727 views • 58 slides
ZooKeeper Wait-free coordination for Internet-scale systems Patrick Hunt and Mahadev (Yahoo!

ZooKeeper Wait-free coordination for Internet-scale systems Patrick Hunt and Mahadev (Yahoo!

ZooKeeper Wait-free coordination for Internet-scale systems Patrick Hunt and Mahadev (Yahoo! Grid) Flavio Junqueira and Benjamin Reed (Yahoo! Research) Internet-scale Challenges Lots of servers, users, data FLP, CAP Mere mortal

271 views • 24 slides
Everybody wants coordination, but nobody wants to be coordinated. Who coordinates the

Everybody wants coordination, but nobody wants to be coordinated. Who coordinates the

Coordination in the field State of the art and lessons learned Alois A. Hirschmugl Austria EULUX WS 1 09.June 07 C O O R D I N A T I O N* O O R D I N A T I O N* C Everybody wants coordination, but nobody wants to be

409 views • 22 slides
Eris: Coordination-Free Consistent Transactions Using In-Network Concurrency Control Jialin Li,

Eris: Coordination-Free Consistent Transactions Using In-Network Concurrency Control Jialin Li,

Eris: Coordination-Free Consistent Transactions Using In-Network Concurrency Control Jialin Li, Ellis Michael, Dan R. K. Ports Web services and applications rely on distributed storage systems Web services and applications rely on distributed

1.31k views • 110 slides
Interval Computations as Why Intervals? Applied Constructive Interval Computations . . . Wiener

Interval Computations as Why Intervals? Applied Constructive Interval Computations . . . Wiener

General Problems of . . . Indirect Measurements How Constructive . . . Applied Constructive . . . Interval Computations as Why Intervals? Applied Constructive Interval Computations . . . Wiener Again Mathematics: from Shanin Standard

385 views • 25 slides
Distributed Hierarchical GPU Parameter Server for Massive Scale Deep Learning Ads Systems

Distributed Hierarchical GPU Parameter Server for Massive Scale Deep Learning Ads Systems

Distributed Hierarchical GPU Parameter Server for Massive Scale Deep Learning Ads Systems Presenter: Weijie Zhao 1 1 Cognitive Computing Lab, Baidu Research Joint work with Deping Xie 2 , Ronglai Jia 2 , Yulei Qian 2 , Ruiquan Ding 3 , Mingming

506 views • 23 slides
Approaching a Platform Migration Approaches to SAS migration and Platform LSF considerations for

Approaching a Platform Migration Approaches to SAS migration and Platform LSF considerations for

Approaching a Platform Migration Approaches to SAS migration and Platform LSF considerations for SAS/Grid Topics Scope Architecture Migration Questions Grid considerations Questions Bell Canada SAS migration 24 months

716 views • 40 slides
Untangling Header Bidding Lore Some myths, some truths, and some hope Waqar Aqeel , Debopam

Untangling Header Bidding Lore Some myths, some truths, and some hope Waqar Aqeel , Debopam

Untangling Header Bidding Lore Some myths, some truths, and some hope Waqar Aqeel , Debopam Bhattacherjee, Balakrishnan Chandrasekaran, Philip Brighten Godfrey, Gregory Laughlin, Bruce M. Maggs, and Ankit Singla 1 How (traditional) Real-Time

442 views • 15 slides
Lecture 10: THE AD-AS MODEL Reference: Chapter 8 LEARNING OBJECTIVES 1.What determines the

Lecture 10: THE AD-AS MODEL Reference: Chapter 8 LEARNING OBJECTIVES 1.What determines the

Lecture 10: THE AD-AS MODEL Reference: Chapter 8 LEARNING OBJECTIVES 1.What determines the shape of the aggregate demand (AD) curve and what factors shift the entire curve. 2.What determines the shape of the aggregate supply (AS) curve, and

289 views • 17 slides
TOOLS FOR BEYOND DIE CODESIGN AND INTEGRATION Honoring Prof. Yoji Kajitani

TOOLS FOR BEYOND DIE CODESIGN AND INTEGRATION Honoring Prof. Yoji Kajitani

ON THE WAY TO PRACTICAL TOOLS FOR BEYOND DIE CODESIGN AND INTEGRATION Honoring Prof. Yoji Kajitani ISPD 2013 Hung-Ming Chen, DEE NCTU, Taiwan Experience I hope the audience to have Honoring Prof. Kajitani By showing

1k views • 56 slides
Advertising and Voter Data in Asymmetric Political Contests Priyanka Sharma Liad Wagman 31

Advertising and Voter Data in Asymmetric Political Contests Priyanka Sharma Liad Wagman 31

Advertising and Voter Data in Asymmetric Political Contests Priyanka Sharma Liad Wagman 31 January 2020 Priyanka Sharma and Liad Wagman Advertising and Voter Data in Asymmetric Political Contests 0 / 17 Motivation According to a recent FTC

653 views • 18 slides
Quar te r ly D&O Claims T r e nds: 2017 Wr ap-Up Ja nua ry 24 th , 2018, 11 AM E a ste

Quar te r ly D&O Claims T r e nds: 2017 Wr ap-Up Ja nua ry 24 th , 2018, 11 AM E a ste

Quar te r ly D&O Claims T r e nds: 2017 Wr ap-Up Ja nua ry 24 th , 2018, 11 AM E a ste rn About Advise n L e a ding the wa y to sma rte r a nd mo re e ffic ie nt risk a nd insura nc e c o mmunitie s, Advise n de live rs: the r

452 views • 22 slides
Smart Pacing for Effective Online Ad Campaign Optimization Jian Xu, Kuang-chih Lee, Wentong Li,

Smart Pacing for Effective Online Ad Campaign Optimization Jian Xu, Kuang-chih Lee, Wentong Li,

Smart Pacing for Effective Online Ad Campaign Optimization Jian Xu, Kuang-chih Lee, Wentong Li, Hang Qi, and Quan Lu Yahoo Inc. August 26, 2015 Jian Xu et al. (Yahoo Inc.) Smart Pacing August 26, 2015 1 / 17 Overview Background 1 Pacing

303 views • 17 slides

More recommend