Multi-core Computing Security: Lecture 43 - Multiprocessor Techniques, Slides of Computer Science

A series of lecture notes on multi-core computing security, specifically focusing on lecture 43: multiprocessor techniques. The content covers various topics such as producer and consumer code, message passing, synchronization with messages, buffering model, and multiprocessor techniques. It also includes exercises and deciphering tasks.

Typology: Slides

2012/2013

Uploaded on 03/28/2013

ekanath
ekanath 🇮🇳

3.8

(4)

76 documents

1 / 10

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Objectives_template
file:///D|/...ary,%20Dr.%20Sanjeev%20K%20Aggrwal%20&%20Dr.%20Rajat%20Moona/Multi-core_Architecture/lecture%2043/43_1.htm[6/14/2012 12:18:10 PM]
Module 22: Multi-core Computing Security
Lecture 43: Multiprocessor Techniques
The Lecture Contains:
Producer and Consumer Code
Message Passing
Synchronization With Messages
Buffering Model
Multiprocessor Techniques
First Job
Decipher
Security Infrastructure
Goals: Confidentiality
Goals: Integrity
Goals: Availability
Policies
Mechanisms
Enforcement Media
Threat and Attack
Information Security
Threat Perception and Cost of Security
Cryptography: кρυptο γρafη(hidden writing)
Cryptography
pf3
pf4
pf5
pf8
pf9
pfa

Partial preview of the text

Download Multi-core Computing Security: Lecture 43 - Multiprocessor Techniques and more Slides Computer Science in PDF only on Docsity!

Module 22: Multi-core Computing Security

Lecture 43: Multiprocessor Techniques

The Lecture Contains:

Producer and Consumer Code

Message Passing

Synchronization With Messages

Buffering Model

Multiprocessor Techniques

First Job

Decipher

Security Infrastructure

Goals: Confidentiality

Goals: Integrity

Goals: Availability

Policies

Mechanisms

Enforcement Media

Threat and Attack

Information Security

Threat Perception and Cost of Security

Cryptography: кρυptο γρafη(hidden writing)

Cryptography

Module 22: Multi-core Computing Security

Lecture 43: Multiprocessor Techniques

Producer and Consumer Code

BUF_Data consume(void) { while (buffer.outptr == buffer.inptr) ; buffer.items[buffer.outptr]; buffer.outptr = (buffer.outptr +1)%BUF_SZ; return (buffer.items[buffer.outptr-1]%BUF_SZ;); }

What is wrong with this code?

Message Passing

Communication channel How to name the channel between two processes? Direct Communication The sender process (P) must know the receiver process (Q) and vice versa P: send(Q, msg) ♦ Q: receive(P, msg); Some versions may have receive(ANY, msg); Indirect Communication Mailboxes must be named rather than the processes. Sender: Create_mailbox(Name, Properties); Sender: Send(Name, msg); Sender: Distroy_mailbox(Name); Receiver: Open_mailbox(Name); Receiver: Receive(mailbox, msg);

Module 22: Multi-core Computing Security

Lecture 43: Multiprocessor Techniques

Multiprocessor Techniques

Depends upon the hardware services. Shared memory machines usually provide shared memory IPC. Distributed memory machines usually provide message passing. On networked machines: Message passing. On multi-core machines: Shared memory and threading.

Multi-core Computing Security

First Job

A quote from Network Security book by Charlie Kaufman, Radia Perlman, Mike Speciner.

Si spy net work, big fedjaw iog link kyxogy

Please decipher it for me.

Hint: Dedication.

Module 22: Multi-core Computing Security

Lecture 43: Multiprocessor Techniques

Decipher

Si spy net work, big fedjaw iog link kyxogy Replace S by T Ti tpy net work, big fedjaw iog link kyxogy Replace i by o To tpy net work,bog fedjaw oog lonk kyxogy Replace p by h To thy net work,bog fedjaw oog lonk kyxogy Replace y by e To the net work,bog fedjaw oog lonk kyxoge Replace n by b To the bet work, bog fedjaw oog lobk kyxoge Replace e by a To the bat work,bog fadjaw oog lobk kyxoge Replace t by d To the bad work,bog fadjaw oog lobk kyxoge Replace k by s To the bad wors, bog fadjaw oog lobs syxoge Replace w by g To the bad gors, bog fadjag oog lobs syxoge Replace o by u To the bad gurs, bog fadjag oug lobs syxuge Replace g by r To the bad gurs, bor fadjag our lobs syxure r byy,bbyf, … To the bad guys, for making our jobs secure

Security Infrastructure

Module 22: Multi-core Computing Security

Lecture 43: Multiprocessor Techniques

Goals: Availability

A most common attack is “denial of service” attack. Attacker does not get the access but can prevent other authorized users getting access as well.

Policies

“What is permitted” For example “only course students can have read access to the these lecture notes” Policies are usually defined by the administrator or owner of the resource.

Mechanisms

Mechanisms are techniques/methods to enforce a policy For example a “attributes” associated with a file can be changed by the owners Mechanism need not even be technical A lost ID card application must be approved by the Dean’s office before a new one is issued. In computer related security, typically procedural mechanisms are used.

Module 22: Multi-core Computing Security

Lecture 43: Multiprocessor Techniques

Enforcement Media

The channel through which the information or a resource access is granted. OS for example. Sometimes the media may not be trustworthy (for example the network) In the security policies and mechanism this aspect has to be taken care of.

Threat and Attack

Threat is a potential violation of security Attack is actual violation. Leakage of information Modification of message while in transit Loss of information Proxy Active and passive attacks

file:///D|/...ry,%20Dr.%20Sanjeev%20K%20Aggrwal%20&%20Dr.%20Rajat%20Moona/Multi-core_Architecture/lecture%2043/43_10.htm[6/14/2012 12:18:12 PM]

Module 22: Multi-core Computing Security

Lecture 43: Multiprocessor Techniques

Mangling of information in a way that unauthorized parties not able to de-mangle. Applications include integrity checking and authentication. Plaintext or cleartext: The message in its original form. Ciphertext: The mangled information.

Cryptography