CS2105 » assignment

Assignment 3: Solution

Ooi Wei Tsang — Tue, 23 Apr 2013 12:04:45 +0000

Here is the solution.

Note that there is much confusion regarding Question 1 Part (l) (are they in the same LAN?). Since this question involves concepts that contradicts what you learnt in CS2105, I have decided to give 2 marks for this question regardless of what your answer is.

You will see an update to your A3 score in IVLE later. (update: 11am 24/4, IVLE updated)

Assignment 2: Solution

Ooi Wei Tsang — Tue, 23 Apr 2013 05:52:37 +0000

Assignment 3: Marks

Ooi Wei Tsang — Mon, 22 Apr 2013 09:43:27 +0000

The marks for Assignment 3 has been loaded into your IVLE gradebook. The solution and explanation will be posted.

If you spot any surprises in your marks. Please don’t panic; send us an email!

Assignment 2: Marks and Comments

Ooi Wei Tsang — Sun, 21 Apr 2013 12:52:48 +0000

I have finished grading Assignment 2. There are several common mistakes that I would like to highlight below:

If a packet is corrupted, you should not make use of the content of the packet! Many of you still uses p.seq even when p.isCorrupted is true.
Sending and receiving the last packet (with a size of zero) should be handled reliably as well. Not handling this may cause the sender or receiver to wait forever.
Some of you send every packet twice. The second parameter of the schedule(task, delay, period) method tells the timer when to first execute the task. If you use a delay of 0, the task (which runs udt.send(p)) is executed immediately. Thus, if you call udt.send(p) in the send() method of RSTSender already, you effectively are sending every packet twice.
To exit gracefully, the sender needs to ensure that any outstanding timer has been cancelled. Many of you did not cancel the timer either in the finally clause or in the catch( ) clause, possibly causing the sender to hang (and continue retransmitting).

The marks for Assignment 2 has been loaded to IVLE. Annotated version of your code has been deposited into your cs2105-z home directory. As usual, if you see any surprises or spot any bug in grading, please email me ASAP.

Assignment 3

Ooi Wei Tsang — Tue, 02 Apr 2013 02:27:32 +0000

Here is your third and final assignment. You may download the PDF version of Assignment 3 here and the packet dump here. You have two weeks to complete this assignment individually. The assignment is due on 15 April, 2013, 6pm. You may hand in a hardcopy to me before the deadline. Some words of advice:

Start early.
Do not copy from each other.

Assignment 2: Common Misconceptions and Questions

Ooi Wei Tsang — Wed, 27 Mar 2013 08:48:31 +0000

We received many questions about the following points:

Should I call send( ) of RDTSender or UDTSender when I retransmit?
Note that when you retransmit, you are not sure if retransmission will be received or not — in other words, retransmission is done UNreliably. If you call send( ) of RDTSender to retransmit, this could work as well, but since this leads to recursion, it is less straightforward and bug prone. Unless you really know what you are doing, I would recommend the most straight forward method.
What should I return if RDT’s recv( ) / send( ) did not receive the expected DataPacket or AckPacket?
This is an invalid question. RDTSender’s send( ) should send the given data reliably. It should not return until it ensures that the given data is received by the receiver. Similarly, RDTReceiver’s recv( ) should return the next chunk of data. It should not return until it can deliver the next chunk of data.
Why did I get XYZ Exception and how should I handle it?
If you get an exception related to object streams, this is very likely caused by having multiple timers active at the same time or multiple send( ) at the same time. Check that you have only one timer at a time, and there is no possibility that two send( ) is called at the same time.

If you get an exception at the sender due to receiver closing of connection after the receiver receives the final packet, this is excepted. You should just catch the exception, clean-up the sender, and return.

If you get a null pointer exception, figure out which pointer is null. Then check that the variable is initialized properly and the variable is accessible within the correct scope. We can’t help much with this one since, unlike the above, this exception is a general programming bug rather than protocol bug.

Assignment 2: Tricky Synchronization Bug Warning and Fix

Ooi Wei Tsang — Fri, 22 Mar 2013 07:04:38 +0000

There is a tricky thread synchronization bug, discovered by Junliang, that could manifest if you create your timer first before sending your DataPacket. The bug causes an IllegalClassException when the receiver reads the DataPacket.

To prevent this bug from biting, you need to either (i) send the DataPacket before setting the timer, and ensure that only one timer is active at a time, or (ii) add the keyword “synchronized” in the send() method of UDTSender as follows:

    /**
     * Given a data packet from RDTSender, this method sends
     * the packet to the receiver by writing it into the
     * object output stream.
     */
    synchronized void send(DataPacket p) throws IOException
    {   
        System.out.println("S: send " + p.seq);
        oos.writeObject(p);
        oos.flush();
    }

You can add this keyword yourself, or copy a new version of UDTSender:

   cp ~sadm/a2/UDTSender.java ~/a2

(after making sure that your local copy of UDTSender.java is writable)

If you are familiar with thread synchronization issues, and would like to know more, read on.

The bug appears if you start the timer first, then send the packet. It is therefore possible that the timer timeout in the middle of transmitting the original DataPacket. Since the given UDTSender.send() code is not in a critical region, the send() execution for the original packet could yield, and the send() execution for the retransmitted packet could start. The receiver gets confused and throws the exception.

Adding the synchronized keyword to UDTSender’s send() ensure that a send() execution always completes before another send() could start.

CS2105 A1 Q&A Schedule (Second Batch)

Ooi Wei Tsang — Sun, 17 Mar 2013 15:28:50 +0000

This page will be continuously updated according to recent sign ups. If you have signed up and your team is not yet listed, please check back periodically for updates.

WHO                  WHEN                          WHERE
A0084383Y            18 March, Monday   4:00pm    AS6 #05-14
A0106572J U096070W   18 March, Monday   4:05pm    AS6 #05-14
A0067407A            18 March, Monday   4:10pm    AS6 #05-14
A0091726B A0091747W  18 March, Monday   4:15pm    AS6 #05-14
A0088776E A0085686L  18 March, Monday   4:20pm    AS6 #05-14

A0069945L A0067387N  18 March, Monday   5:15pm    AS6 #05-14
A0081146L A0085736U  18 March, Monday   5:20pm    AS6 #05-14
A0083648U U0909018U  18 March, Monday   5:25pm    AS6 #05-14	

A0091621M A0091834B  19 March, Tuesday  3:00pm    AS6 #05-14
A0088828J A0084644Y  19 March, Tuesday  3:05pm    AS6 #05-14	
A0106855B ...	     19 March, Tuesday  3:10pm    AS6 #05-14

Assignment 1 Q&A Session (Second Batch)

Ooi Wei Tsang — Sun, 17 Mar 2013 03:35:01 +0000

For those who did not sign up, did not show up, or were away last week and missed the Q&A session, please sign up for your Q&A session at http://doodle.com/kmuez67yr2cnawv6

Note that the week of 18 March is the last week where Q&A session is organized. We will finalize the marks for Assignment 1 after this and any no-show will receive 0 for Q&A.

Assignment 1 Q&A Schedule (Updated)

Ooi Wei Tsang — Wed, 13 Mar 2013 01:53:57 +0000

Update: 14 March 4:40PM. The venue for Friday’s session (after 11am) has been changed from DR5 to DR9 (COM2 04-06). The venue for Friday 10am – 11am session remains the same.

Please see below for the Q&A schedule. Please be punctual. We will do our best to keep to the schedule. Please excuse us for not having time for pleasantries and small talks during the Q&A session :)

Please report any bug in the schedule.

U090328L U093557R	14 Mar, 2013	12:00 PM	AS6 05-14
A0091038J A0087537R	14 Mar, 2013	12:05 PM	AS6 05-14
A0081254L	        14 Mar, 2013	12:10 PM	AS6 05-14
A0067407A A0106834H	14 Mar, 2013	12:15 PM	AS6 05-14
A0091646Y A0091683X	14 Mar, 2013	12:20 PM	AS6 05-14
A0087580W A0092051R	14 Mar, 2013	12:30 PM	AS6 05-14
A0091621M  A0091834B	14 Mar, 2013	12:35 PM	AS6 05-14
A0106661L A0104785B	14 Mar, 2013	12:40 PM	AS6 05-14
A0088447N A0088454U	14 Mar, 2013	12:45 PM	AS6 05-14
A0088271Y A0088816N	14 Mar, 2013	12:50 PM	AS6 05-14
A0090372P A0090373S	14 Mar, 2013	1:00 PM	        AS6 05-14
A0091901L  A0087652W	14 Mar, 2013	1:05 PM	        AS6 05-14
A0087510J	        14 Mar, 2013	1:10 PM	        AS6 05-14
A0091798J A0091784U	14 Mar, 2013	1:15 PM	        AS6 05-14
A0091657W A0105518L	14 Mar, 2013	1:20 PM	        AS6 05-14
U096854J A0087409W	14 Mar, 2013	1:30 PM	        AS6 05-14
A0084960X A0085356X	14 Mar, 2013	1:35 PM	        AS6 05-14
U095630E U095608J	14 Mar, 2013	1:40 PM	        AS6 05-14
A0091747W A0091726B	14 Mar, 2013	1:45 PM	        AS6 05-14
U096891J A0085264B	14 Mar, 2013	2:00 PM	        AS6 05-14
A0087666L A0090941E	14 Mar, 2013	2:05 PM	        AS6 05-14
A0091903H  A0091894N	14 Mar, 2013	2:10 PM	        AS6 05-14
A0087876E	        14 Mar, 2013	2:15 PM	        AS6 05-14
A0091725A  A0091656X	14 Mar, 2013	2:20 PM	        AS6 05-14
A0106409L	        14 Mar, 2013	2:30 PM	        AS6 05-14
A0087745N A0091024U	14 Mar, 2013	2:35 PM	        AS6 05-14
A0073817Y U097902X	14 Mar, 2013	2:40 PM	        AS6 05-14
A0105507N A0105610Y	14 Mar, 2013	2:45 PM	        AS6 05-14
A0096857H	        14 Mar, 2013	4:00 PM	        DR5, AS6 02-08
A0091694U A0091658U	14 Mar, 2013	4:05 PM	        DR5, AS6 02-08
A0078120N A0085897A	14 Mar, 2013	4:10 PM	        DR5, AS6 02-08
A0081075J A0083431M	14 Mar, 2013	4:15 PM	        DR5, AS6 02-08
A0090585 A0087914U	14 Mar, 2013	4:20 PM	        DR5, AS6 02-08
U096752MU096763W	14 Mar, 2013	4:30 PM	        DR5, AS6 02-08
A0075019H A0086514A	14 Mar, 2013	4:35 PM	        DR5, AS6 02-08
A0091723H,A0091736Y	14 Mar, 2013	4:40 PM	        DR5, AS6 02-08
A0077901A A0077826R	14 Mar, 2013	4:45 PM	        DR5, AS6 02-08
A0003900B A0083521M	14 Mar, 2013	4:50 PM	        DR5, AS6 02-08
A0106829Y	        14 Mar, 2013	5:00 PM	        DR5, AS6 02-08
U096785L U093020H	14 Mar, 2013	5:05 PM	        DR5, AS6 02-08
A0078037B A0092000A	14 Mar, 2013	5:10 PM	        DR5, AS6 02-08
A0088324B A0087568J	14 Mar, 2013	5:15 PM	        DR5, AS6 02-08
A0091810M A0091811L	14 Mar, 2013	5:20 PM	        DR5, AS6 02-08
U090269J U096794W	14 Mar, 2013	5:25 PM	        DR5, AS6 02-08
A0088320J A0088574M	15 Mar, 2013	10:00 AM	AS6 05-14	
A0067402M A0074008M	15 Mar, 2013	10:05 AM	AS6 05-14	
A0106715L A0088498A	15 Mar, 2013	10:10 AM	AS6 05-14	
A0086931X A0088827L	15 Mar, 2013	10:15 AM	AS6 05-14	
A0075068Y A0075241M	15 Mar, 2013	10:20 AM	AS6 05-14	
A0084667N A0081241U	15 Mar, 2013	10:30 AM	AS6 05-14	
A0077979 U0906806	15 Mar, 2013	10:35 AM	AS6 05-14	
A0084692U A0084444A	15 Mar, 2013	11:15 AM	DR9, COM2 04-06
A0091741H A0075021W	15 Mar, 2013	11:20 AM	DR9, COM2 04-06
A0072320X A0091013X	15 Mar, 2013	11:25 AM	DR9, COM2 04-06
A0087974H	        15 Mar, 2013	11:30 AM	DR9, COM2 04-06
A0085314J A0087084X	15 Mar, 2013	11:35 AM	DR9, COM2 04-06
A0087752 A0087947H	15 Mar, 2013	11:40 AM	DR9, COM2 04-06
A0085380A A0084959H	15 Mar, 2013	11:45 AM	DR9, COM2 04-06
U096783M U096809L	15 Mar, 2013	11:50 AM	DR9, COM2 04-06
A0097710A	        15 Mar, 2013	12:00 PM	DR9, COM2 04-06
A0091846W A0091908X	15 Mar, 2013	12:05 PM	DR9, COM2 04-06
A0077954M A0078063A	15 Mar, 2013	12:10 PM	DR9, COM2 04-06
A0073763 A0074030	15 Mar, 2013	1:00 PM	        DR9, COM2 04-06
A0086772R A0087171A	15 Mar, 2013	1:05 PM	        DR9, COM2 04-06
U096834R U096867N	15 Mar, 2013	1:10 PM	        DR9, COM2 04-06
A0088441B A0083545B	15 Mar, 2013	1:15 PM	        DR9, COM2 04-06
A0083580A U086470N	15 Mar, 2013	1:20 PM	        DR9, COM2 04-06
A0075082H A0073758R	15 Mar, 2013	1:30 PM	        DR9, COM2 04-06
A0075054J A0075059Y	15 Mar, 2013	1:35 PM	        DR9, COM2 04-06
A0097758E A0097787B	15 Mar, 2013	1:40 PM	        DR9, COM2 04-06
A0106425N A0106429H	15 Mar, 2013	1:45 PM	        DR9, COM2 04-06
A0085594N A0084053L	15 Mar, 2013	1:50 PM	        DR9, COM2 04-06