20CS02002 20CS01014 20CS01061 CC Assignment

Cloud Computing
Assignment
By Thadicherla Hrishith(20CS02002)
Mallela Sathvik(20CS01061)
P.R.R Sumiran Adithya(20CS01014)
Project Statement:
The project requires implementation of Lamport’s Mutual exclusion
using logical clock. The project has to be implemented via sockets
because you need to communicate between 3 devices. The critical
section can be assumed to be access to a file on any one of the 3
machines. There need to be two threads per device. One thread shall
create local event and send event; the send event should only be for
mutual exclusion request. The other thread need to act as port listener
and hence, receiver thread. The only catch here is to synchronise
between sender and receiver thread to update the logical clock. The
implementation needs to be done only in C++ or Java.
Approach to Solution:
The first part of this problem is to create the sender and listener
threads. My approach to this was that don’t need to keep the sender
thread always on. We can just create the sender thread whenever are
are making the request and just send it to the 2 other clients. However
the listener thread needs to be active all the time. So we activate it
when we first take in the ip and port.
The second part of this problem is how do we access the file hosted
on one system from another system.
One approach is to ask the process whose system is hosting it to edit

it, but it seems wrong because we are giving one process on the
system direct access to the data of the other process in the system.
Also, the file access needs to be fair, we can understand why the code
to access the file shouldn’t be in the main code.
Other approach is creating another process on this same system which
has edit access to this file, attach a socket to it and then let all the
systems access this process through mutual exclusion. The access is
only possible since we attached a socket to it.
File called CriticalSection.java is the code which has access to the
file.
Next issue is synchronisation of clocks , queues and replies.
The way we can do that is by usage of locks. This way we can make
sure that only one thread sender or listener is accessing these sections
at the same time.
We will create a timeStamp Lock, PriorityqueueLock and replyLock
and surround it whenever we need to update timeStamp or insert into
Priorityqueue or reset or increment the replies.
That’s pretty much the end of our issues outside the logic of the
algorithm . All we need to do is see the issues inside Lamport’s
Mutual Exclusion.
The Algorithm as in the text book Advanced Operating Systems by

Shivratri.
Requesting the critical section:
1. When a site Si wants to enter the CS, it sends a
REQUEST(ts{i}, i) message to all the sites in its request set
R_{i} and places the request on request queue. ( (ts{i}, i) is the
timestamp of the request.)
2. When a site S_{j} receives the REQUEST(ts{i}, i) message
from site S_{i} it returns a timestamped REPLY message S_{i}
and places site S_{i}' s request on request queues
Executing the critical section: Site S, enters the CS when the two
following conditions hold.
[L1:] S_{i} has received a message with timestamp larger than
(ts_{i}, i) from all other sites.
[L2:] S_{i} ‘s request is at the top of request queue
Releasing the critical section:
3. Site S_{i} upon exiting the CS, removes its request from the top
of its request queue and sends a timestamped RELEASE message
to all the sites in its request set.
4. When a site S_{j} receives a RELEASE message from site

S_{i} it removes S_{i}’ s request from its request queue.
When a site removes a request from its request queue, its own
request may come at the top of the queue, enabling it to enter the
CS. The algorithm executes CS requests in the increasing
order of timestamps.
The messages are send in the format :

TypeofMessage_Timestamp@Ip:Port
From this we can easily decode the message by using split in java.
So the listener thread receives all kinds of messages REQUEST,
REPLY and RESPONSE.
Sender thread can create the Event and send requests.

The main issue we have after synchronising everything with locks
is
Accessing the critical section:
The way we implement that is as follows:
1. If we are on the top of queue then it is straight forward. We
should access the critical section .So other systems with check
their queues send the replies and it is done.
2. If we are not on top of queue then it becomes tricky. In that case
where some other pc is accessing CS. When it sends the release
message. What we need to do is pop the queue. Now you have
two cases.
1. If after popping we are on top of queue then we are not
supposed to do anything. We have to wait for the one who
just released the CS to send a reply and the other pc should
check the queue and send the reply
2. If after popping we are not on top of queue then we are
supposed to send a reply to whoever is on the top of the
queue.
After this the logic is pretty straight forward and explained in the code
comments.
Here is an example of how this runs:

We are running this in localhost with ports 6000, 6001,6002
The above example is of 6000 giving request first and 6001 giving
immediate request next.
Here is a screenshot of timestamps of all the events that happened:

20CS02002 20CS01014 20CS01061 CC Assignment

Uploaded by

Copyright:

Available Formats

20CS02002 20CS01014 20CS01061 CC Assignment

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

20CS02002 20CS01014 20CS01061 CC Assignment

Uploaded by

Copyright:

Available Formats

Cloud Computing

One approach is to ask the process whose system is hosting it to edit

The Algorithm as in the text book Advanced Operating Systems by

[L2:] S_{i} ‘s request is at the top of request queue

Releasing the critical section:

4. When a site S_{j} receives a RELEASE message from site

The messages are send in the format :

Sender thread can create the Event and send requests.

Here is an example of how this runs:

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.