1. 08 April 2004 1 PPS - Groupe de Travail en Concurrence The probabilistic asynchronous -calculus Catuscia Palamidessi, INRIA Futurs, France

2. 08 April 2004PPS - Groupe de Travail en Concurrence 2 Motivations Main motivation behind the development of the probabilistic asynchronous  -calculus: To use it as an intermediate language for the fully distributed implementation of the  -calculus with mixed choice Plan of the talk The  -calculus, Operational semantics Expressive power The probabilistic asynchronous  -calculus Probabilistic automata, operational semantics Distributed implementation Encoding the  -calculus into the probabilistic asynchronous  -calculus

3. 08 April 2004PPS - Groupe de Travail en Concurrence 3 The  -calculus Proposed by [Milner, Parrow, Walker ‘92] as a formal language to reason about concurrent systems Concurrent: several processes running in parallel Asynchronous: every process proceeds at its own speed Synchonous communication: aka handshaking Mixed guarded choice: input and output guards like in CSP and CCS. The implementation of guarded choice is aka the binary interaction problem Dynamic generation of communication channels Scope extrusion: a channel name can be communicated and its scope extended to include the recipient xy z z z R Q P

4. 08 April 2004PPS - Groupe de Travail en Concurrence 4  : the  -calculus (w/ mixed choice) Syntax g ::= x(y) | x^y |  prefixes (input, output, silent) P ::=  i g i. P i mixed guarded choice | P | P parallel | (x) P new name |rec A P recursion | A procedure name

5. 08 April 2004PPS - Groupe de Travail en Concurrence 5 Operational semantics Transition system P -a  Q Rules Choice  i g i. P i –g i  P i P -x^y  P’ Open ___________________ (y) P -x^(y)  P’

6. 08 April 2004PPS - Groupe de Travail en Concurrence 6 Operational semantics Rules (continued) P -x(y)  P’ Q -x^z  Q’ Com ________________________ P | Q -   P’ [z/y] | Q’ P -x(y)  P’ Q -x^(z)  Q’ Close _________________________ P | Q -   (z) (P’ [z/y] | Q’) P -g  P’ Par _________________ fn(Q) and bn(g) disjoint Q | P -g  Q | P

7. 08 April 2004PPS - Groupe de Travail en Concurrence 7 Features which make  very expressive - and cause difficulty in its distributed implementation (Mixed) Guarded choice Symmetric solution to certain distributed problems involving distributed agreement Link mobility Network reconfiguration It allows expressing HO (e.g. calculus) in a natural way In combination with guarded choice, it allows solving more distributed problems than those solvable by guarded choice alone

8. 08 April 2004PPS - Groupe de Travail en Concurrence 8 Example of distributed agreement: The leader election problem in a symmetric network Two symmetric processes must elect one of them as the leader In a finite amount of time The two processes must agree A symmetric and fully distributed solution in  using guarded choice: x.P wins + y^.P loses | y.Q wins + x^.Q loses The expressive power of   P loses | Q wins PQ y x  P wins | Q loses

9. 08 April 2004PPS - Groupe de Travail en Concurrence 9 Example of a network where the leader election problem cannot be solved by guarded choice alone For the following network there is no (fully distributed and symmetric) solution in CCS, or in CSP

10. 08 April 2004PPS - Groupe de Travail en Concurrence 10 A solution to the leader election problem in  looser winner looser winnerlooserwinner

11. 08 April 2004PPS - Groupe de Travail en Concurrence 11 Approaches to the implementation of guarded choice in literature [Parrow and Sjodin 92], [Knabe 93], [Tsai and Bagrodia 94]: asymmetric solution based on introducing an order on processes Other asymmetric solutions based on differentiating the initial state Plenty of centralized solutions [Joung and Smolka 98] proposed a randomized solution to the multiway interaction problem, but it works only under an assumption of partial synchrony among processes Our solution is the first to be fully distributed, symmetric, and using no synchronous hypotheses.

12. 08 April 2004PPS - Groupe de Travail en Concurrence 12 State of the art in  Formalisms able to express distributed agreement are difficult to implement in a distributed fashion For this reason, the field has evolved towards variants of  which retain mobility, but have no guarded choice One example of such variant is the asynchronous  calculus proposed by [Honda- Tokoro’91, Boudol, ’92] (Asynchronous = Asynchronous communication)

13. 08 April 2004PPS - Groupe de Travail en Concurrence 13  a : the Asynchonous  Version of [Amadio, Castellani, Sangiorgi ’97] Syntax g ::= x(y) |  prefixes P ::=  i g i. P i input guarded choice |x^y output action | P | P parallel | (x) P new name |rec A P recursion | A procedure name

14. 08 April 2004PPS - Groupe de Travail en Concurrence 14 Characteristics of  a Asynchronous communication: we can’t write a continuation after an output, i.e. no x^y.P, but only x^y | P so P will proceed without waiting for the actual delivery of the message Input-guarded choice: only input prefixes are allowed in a choice. Note: the original asynchronous  calculus did not contain a choice construct. However the version presented here was shown by [Nestmann and Pierce, ’96] to be equivalent to the original asynchronous  calculus It can be implemented in a fully distributed fashion (see for instance Odersky’s group’s project PiLib)

15. 08 April 2004PPS - Groupe de Travail en Concurrence 15 Towards a fully distributed implementation of  The results of previous pages show that a fully distributed implementation of  must necessarily be randomized A two-steps approach:  probabilistic asynchronous  distributed machine [[ ]] > Advantages: the correctness proof is easier since [[ ]] (which is the difficult part of the implementation) is between two similar languages

16. 08 April 2004PPS - Groupe de Travail en Concurrence 16  pa : the Probabilistic Asynchonous  Syntax g ::= x(y) |  prefixes P ::=  i p i g i. P i pr. inp. guard. choice  i p i = 1 |x^y output action | P | P parallel | (x) P new name |rec A P recursion | A procedure name

17. 08 April 2004PPS - Groupe de Travail en Concurrence 17 1/2 1/3 2/3 1/2 1/3 2/3 1/2 1/3 2/3 The operational semantics of  pa Based on the Probabilistic Automata of Segala and Lynch Distinction between nondeterministic behavior (choice of the scheduler) and probabilistic behavior (choice of the process) Scheduling Policy: The scheduler chooses the group of transitions Execution: The process chooses probabilistically the transition within the group

18. 08 April 2004PPS - Groupe de Travail en Concurrence 18 The operational semantics of  pa Representation of a group of transition P { --g i -> p i P i } i Rules Choice  i p i g i. P i {--g i -> p i P i } i P {--g i -> p i P i } i Par ____________________ Q | P {--g i -> p i Q | P i } i

19. 08 April 2004PPS - Groupe de Travail en Concurrence 19 The operational semantics of  pa Rules (continued) P {--x i (y i )-> p i P i } i Q {--x^z-> 1 Q’ } i Com____________________________________ P | Q {--  -> p i P i [z/y i ] | Q’ } x i =x U { --x i (y i )-> p i P i | Q } x i =/=x P {--x i (y i )-> p i P i } i Res___________________ q i renormalized (x) P { --x i (y i )-> q i (x) P i } x i =/= x

20. 08 April 2004PPS - Groupe de Travail en Concurrence 20 Implementation of  pa Compilation in Java > :  pa  Java Distributed > = >. start(); >.start(); Compositional > = > jop > for all op Channels are one-position buffers with test-and-set (synchronized) methods for input and output

21. 08 April 2004PPS - Groupe de Travail en Concurrence 21 Encoding  into  pa [[ ]] :    pa Fully distributed [[ P | Q ]] = [[ P ]] | [[ Q ]] Preserves the communication structure [[ P  ]] = [[ P ]]  Compositional [[ P op Q ]] = C op [ [[ P ]], [[ Q ]] ] Correct wrt a notion of probabilistic testing semantics P must O iff [[ P ]] must [[ O ]] with prob 1

22. 08 April 2004PPS - Groupe de Travail en Concurrence 22 Encoding  into  pa Idea (from an idea in [Nestmann’97]): Every mixed choice is translated into a parallel comp. of processes corresponding to the branches, plus a lock f The input processes compete for acquiring both its own lock and the lock of the partner The input process which succeeds first, establishes the communication. The other alternatives are discarded The problem is reduced to a generalized dining philosophers problem where each fork (lock) can be adjacent to more than two philosophers Further, we can reduce the generalized DP to the classic case, and then apply the randomized algorithm of Lehmann and Rabin for the dining philosophers P Q R Pi Qi Ri f f f S R’i f Si

23. 08 April 2004PPS - Groupe de Travail en Concurrence 23 Dining Philosophers: classic case Each fork is shared by exactly two philosophers

24. 08 April 2004PPS - Groupe de Travail en Concurrence 24 The algorithm of Lehmann and Rabin 1. think; 2. choose probabilistically first_fork in {left,right}; 3. if not taken(first_fork) then take(first_fork) else goto 3; 4. if not taken(second_fork) then take(second_fork); else { release(first_fork); goto 2 } 5. eat; 6. release(second_fork); 7. release(first_fork); 8. goto 1

25. 08 April 2004PPS - Groupe de Travail en Concurrence 25 Dining Philosophers: generalized case Each fork can be shared by more than two philosophers The classical algorithm of Lehmann and Rabin, as it is, does not work in the generalized case. However, we can transform it into the classic case Transformation into the classic case: each fork is initially associated with a token. Each phil needs to acquire a token in order to participate to the competition. The competing phils determine a set of subgraphs in which each subgraph contains at most one cycle

26. 08 April 2004PPS - Groupe de Travail en Concurrence 26 Generalized philosophers Another problem we had to face: the solution of Lehmann and Rabin works only for fair schedulers, while  pa does not provide any guarantee of fairness Fortunately, it turns out that the fairness is required only in order to avoid a busy-waiting livelock at instruction 3. If we replace busy-waiting with suspension, then the algorithm works for any scheduler This result was achieved independently also by [Duflot, Fribourg, Picarronny 02].

27. 08 April 2004PPS - Groupe de Travail en Concurrence 27 1. think; 2. choose probabilistically first_fork in {left,right}; 3. if not taken(first_fork) then take(first_fork) else wait; 4. if not taken(second_fork) then take(second_fork); else { release(first_fork); goto 2 } 5. eat; 6. release(second_fork); 7. release(first_fork); 8. goto 1 1. think; 2. choose probabilistically first_fork in {left,right}; 3. if not taken(first_fork) then take(first_fork) else goto 3; 4. if not taken(second_fork) then take(second_fork); else { release(first_fork); goto 2 } 5. Eat; 6. release(second_fork); 7. release(first_fork); 8. goto 1 The algorithm of Lehmann and Rabin Modified so to avoid the need for fairness The algorithm of Lehmann and Rabin

28. 08 April 2004PPS - Groupe de Travail en Concurrence 28 Conclusion We have provided an encoding of the  calculus into a probabilistic version of its asynchronous fragment fully distributed compositional correct wrt a notion of testing semantics Advantages: high-level solutions to distributed algorithms Easier to prove correct (no reasoning about randomization required)