P Systems and Distributed Computing

43
On P Systems and Distributed Computing Apostolos Syropoulos Greek Molecular Computing Group Xanthi, GREECE [email protected] – p. 1/15

Transcript of P Systems and Distributed Computing

Page 1: P Systems and Distributed Computing

On P Systems and Distributed

Computing

Apostolos Syropoulos

Greek Molecular Computing Group

Xanthi, GREECE

[email protected]

– p. 1/15

Page 2: P Systems and Distributed Computing

P Systems and Distributing Computing

Is there any relationship between P systemsand distributed computing?

– p. 2/15

Page 3: P Systems and Distributed Computing

P Systems and Distributing Computing

Is there any relationship between P systemsand distributed computing?

Distributed simulations of P Systems haveshown that such a link exists.

– p. 2/15

Page 4: P Systems and Distributed Computing

P Systems and Distributing Computing

Is there any relationship between P systemsand distributed computing?

Distributed simulations of P Systems haveshown that such a link exists.

Our work aims to make more “formal” thisrelationship.

– p. 2/15

Page 5: P Systems and Distributed Computing

What is Distributed Computing?

Distributed computing is the process ofrunning a single computational task on morethan one distinct computers.

– p. 3/15

Page 6: P Systems and Distributed Computing

What is Distributed Computing?

Distributed computing is the process ofrunning a single computational task on morethan one distinct computers.

Distributed Computing 6= Cluster Computing!

– p. 3/15

Page 7: P Systems and Distributed Computing

What is Distributed Computing?

Distributed computing is the process ofrunning a single computational task on morethan one distinct computers.

Distributed Computing 6= Cluster Computing!

Distributed computing can utilizecomputational resources that would beotherwise unused.

– p. 3/15

Page 8: P Systems and Distributed Computing

What is Distributed Computing?

Distributed computing is the process ofrunning a single computational task on morethan one distinct computers.

Distributed Computing 6= Cluster Computing!

Distributed computing can utilizecomputational resources that would beotherwise unused.

SETI@home (SETI at home) is a typicaldistributed computing project.

– p. 3/15

Page 9: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp.

Network Structure.

Hierarchical

Star

Ring

Multiaccess

Hybrid

– p. 4/15

Page 10: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp.

Network Structure.

Hierarchical

Star

Ring

Multiaccess

Hybrid

Programming Tools

– p. 4/15

Page 11: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp.

Network Structure.

Hierarchical

Star

Ring

Multiaccess

Hybrid

Programming Tools

File “Systems”

– p. 4/15

Page 12: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp.

Network Structure.

Hierarchical

Star

Ring

Multiaccess

Hybrid

Programming Tools

File “Systems”

– p. 4/15

Page 13: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp.

Network Structure.

Hierarchical

Star

Ring

Multiaccess

Hybrid

Programming Tools

File “Systems”

– p. 4/15

Page 14: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp.

Network Structure.

Hierarchical

Star

Ring

Multiaccess

Hybrid

Programming Tools

File “Systems”

– p. 4/15

Page 15: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp.

Network Structure.

Hierarchical

Star

Ring

Multiaccess

Hybrid

Programming Tools

File “Systems”

– p. 4/15

Page 16: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp.

Network Structure.

Hierarchical

Star

Ring

Multiaccess

Hybrid

Programming Tools

File “Systems”

– p. 4/15

Page 17: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp. cont.

Coordination

– p. 5/15

Page 18: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp. cont.

Coordination

Distributed Architectures

client-server architecture

peer-to-peer architecture

– p. 5/15

Page 19: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp. cont.

Coordination

Distributed Architectures

client-server architecture

peer-to-peer architecture

– p. 5/15

Page 20: P Systems and Distributed Computing

Basic Characteristics of Distr. Comp. cont.

Coordination

Distributed Architectures

client-server architecture

peer-to-peer architecture

– p. 5/15

Page 21: P Systems and Distributed Computing

Network Structure

A hierarchical network can be easilyrepresented as a membrane structure.

– p. 6/15

Page 22: P Systems and Distributed Computing

Network Structure

A hierarchical network can be easilyrepresented as a membrane structure.

A star network structure is a restricted form ofa tree structure; ergo nodes can beconsidered as compartments that are placedinside a skin membrane.

– p. 6/15

Page 23: P Systems and Distributed Computing

Network Structure

A hierarchical network can be easilyrepresented as a membrane structure.

A star network structure is a restricted form ofa tree structure; ergo nodes can beconsidered as compartments that are placedinside a skin membrane.

But in general a network structure is actuallya graph.

– p. 6/15

Page 24: P Systems and Distributed Computing

Network Structure

A hierarchical network can be easilyrepresented as a membrane structure.

A star network structure is a restricted form ofa tree structure; ergo nodes can beconsidered as compartments that are placedinside a skin membrane.

But in general a network structure is actuallya graph.

Only P systems with a graph membranestructure are suitable to describe the networkstructure of any system.

– p. 6/15

Page 25: P Systems and Distributed Computing

Graph-structured P Systems

The membrane structure is replaced by arelation in {1, . . . ,m} × {1, . . . ,m}, describingthe network structure of the system

– p. 7/15

Page 26: P Systems and Distributed Computing

Graph-structured P Systems

The membrane structure is replaced by arelation in {1, . . . ,m} × {1, . . . ,m}, describingthe network structure of the system

The new symbol “toj” should be used todirectly place an object from the hostcompartment to compartment j.

– p. 7/15

Page 27: P Systems and Distributed Computing

Graph-structured P Systems

The membrane structure is replaced by arelation in {1, . . . ,m} × {1, . . . ,m}, describingthe network structure of the system

The new symbol “toj” should be used todirectly place an object from the hostcompartment to compartment j.

The rule is applicable only if (i, j) ∈ gm or(j, i) ∈ gm, where gm is the relation describingthe network structure.

– p. 7/15

Page 28: P Systems and Distributed Computing

P Systems and Distr. Progr.

P systems have at least the computationalpower of Turing machines.

– p. 8/15

Page 29: P Systems and Distributed Computing

P Systems and Distr. Progr.

P systems have at least the computationalpower of Turing machines.

In addition, it is possible to encode anygeneral recursive function as P system.

– p. 8/15

Page 30: P Systems and Distributed Computing

P Systems and Distr. Progr.

P systems have at least the computationalpower of Turing machines.

In addition, it is possible to encode anygeneral recursive function as P system.

Thus, we can implement “simple” functionalprogramming languages on P hardware.

– p. 8/15

Page 31: P Systems and Distributed Computing

On P hardware

Designing a new kind of hardware is not anyeasy process.

– p. 9/15

Page 32: P Systems and Distributed Computing

On P hardware

Designing a new kind of hardware is not anyeasy process.

First, we need to design the hardware.

– p. 9/15

Page 33: P Systems and Distributed Computing

On P hardware

Designing a new kind of hardware is not anyeasy process.

First, we need to design the hardware.

Second, we need to implement it in silico or toimplement it in the form of a virtual machine.

– p. 9/15

Page 34: P Systems and Distributed Computing

P processors

Each processor implements the functionalityof a compartment of a particular P system.

– p. 10/15

Page 35: P Systems and Distributed Computing

P processors

Each processor implements the functionalityof a compartment of a particular P system.

Each P-processor must be able tocommunicate with other P-processors.

– p. 10/15

Page 36: P Systems and Distributed Computing

P processors

Each processor implements the functionalityof a compartment of a particular P system.

Each P-processor must be able tocommunicate with other P-processors.

Thus, P-processors will form a network.

– p. 10/15

Page 37: P Systems and Distributed Computing

P Processor Intructions

send d, i Send all “d’s” to processor “i”

isalive i Is P-processor i is alive

intro o, n, i Introduce to the ith compartment

n copies of o

addrule R, i Associate rule R with processor i

delrule R, i Disassociate rule R from

processor i

– p. 11/15

Page 38: P Systems and Distributed Computing

P Processor Intructions cont.

replace o, o′, i Replace each occurrence of o in

compartment i with o′

delete o, i Delete each occurrence of o in

compartment i

ifempty o, i If the compartment i is empty,

then place a copy of o at it

halt n, i Halt processor i after n cycles

ihalt i Immediately halt processor i

– p. 12/15

Page 39: P Systems and Distributed Computing

P Processor Intructions cont.

joinsys i Processor i joins the system

exec Start execution

noop A do-nothing instuction

– p. 13/15

Page 40: P Systems and Distributed Computing

Open Problems

How do we specify the specify the possitionof a P-processor that joins a system?

– p. 14/15

Page 41: P Systems and Distributed Computing

Open Problems

How do we specify the specify the possitionof a P-processor that joins a system?

How do we enumerate the variousP-processors?

– p. 14/15

Page 42: P Systems and Distributed Computing

Open Problems

How do we specify the specify the possitionof a P-processor that joins a system?

How do we enumerate the variousP-processors?

How do we implement the maximalparallelism principle?

– p. 14/15

Page 43: P Systems and Distributed Computing

That’s all!

Thank you very much for your attention!

– p. 15/15