• 0

Bytes with TCP socket and bit manipulation


Question

I'm taking a computer communication summer class and this is our 2nd and last program project. The protocol for the assignment can be found here.

I've already done one project that involves a client and server connecting and playing a game of guess the number with strings being sent back and forth. For this project though we have to manipulate individual bits of bytes and send only bytes across the socket connection. How do you do this in Java? I haven't been able to find anything about forming bytes with particular bits and sending just bytes across the connection.

11 answers to this question

Recommended Posts

  • 0

I'm not sure what your project assignment is doing (I dont' have time to read it). But from my experience in Java, to read binary data you use InputStream or OutputStream to read and write binary data.

Here's an example:

Let's assume "sock" is your socket connection.

InputStream in = sock.getInputStream();
while (in.read() != -1)
{
     //do whatever you need to do with it
}

You can also look at sun's java api for more info (search for inputstream).

Hope this helps!

  • 0

I've read some of the things on the input and output streams and i'll be definitely be using them. My problem comes after I've received the bytes. I have to use certain control bytes as notifiers for things:

00000001- REQuest connection.

00000010 - CONnection Granted.

ssss0011 - FILe name packet (sequence number sss = 000).

ssss0100 - DATa Packet, where sss is the sequence number of the packet.

ssss0101 - Short Data Packet(SDP), where sss is the sequence number of the packet.

ssss0110 - ACKnowledgment packet, where sss is the sequence number of the packet expected next.

ssss 0111 - list request packet (sequence number sss = 000).

0000 1111 - TERmination (a simple two-way handshake).

After I've received the bytes I don't understand how to interpret the different bits of the bytes to know what I'm dealing with.

  • 0

You can AND them (note this is C-style code)

if(byte == 0x01) // REQuest connection
{
}
else if(byte == 0x02) // CONnection granted
{
}
else if(byte & 0x0F == 0x03) // FILe name packet
{
    sequence = (byte & 0xF0) >> 4;
}

  • 0

Thx for the info Andareed!

I've taken computer structures so I understand the idea of And'ing and Or'ing things but I don't quite understand the hex strings your using. Any extra tips on those? I tried google'ing some stuff but it didn't really help. Also, what exactly does the >> do.

Edit:

Since hex is 16 bits I'm guessing the first digit is the first 4 bits and when you check for 0x0F you're testing to see if the first 3 bits are all 1's to know it's a file packet and similarly for the others. The >> still confuses me though.

Edited by Little Moe
  • 0

I've run into another problem. When sending a packet with file information one of the bytes has to be a one-byte unsigned binary value. I've found a whole lot of stuff about doing unsigned bytes in Java by using ints and AND'ing stuff with 0x0FF but my I think my problem comes with actually sending the byte.

My teacher expects our clients and servers to run against his C clients and servers. He gave us an example exe of each and whenever I run my client against his it says "Number of char in file name should be between 1 and 124 - LEN field = 0" like it's not getting my byte at all. I'm thinking it's either the byte I'm using or something special in the OutputStream that might be messing up my byte? I know there is no difference between the bits of a signed or unsigned byte but something is going wrong. Does anyone have an example of sending a byte like he is asking or maybe if I'm thinking of this wrong.

  • 0

I ended up finishing the project. My biggest problem starting off was I was trying to send bytes individually by just writing the bytes one at a time. My teacher was expecting the bytes one right after another. I had to send a byte array for him to get everything in a packet that I wanted. Couldn't just test one byte at a time. Had to test things just by hardcoding byte array values to see what happened.

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Posts

    • Chrome is now faster than ever and Google explains how it did it by Sayan Sen Back in June last year, Google touted some great performance improvements for Chrome and shared a blog post explaining in detail how it managed to achieve them. Today, almost exactly a year later, the search giant is back again with another such post as it continues to make performance gains in its Chromium browser. Interestingly, Google is not the only one to make such claims in recent times. Microsoft also highlighted recently how Edge was getting significantly faster. Both Microsoft and Google have cited the Speedometer 3.0 benchmark to test. We recently measured browsing performance during our T-Force DDR5-7200 RAM review, also using Speedometer 3.0. In its blog post, Google says that the development team made significant improvements to memory management and caching. This includes some redesigning effort of the memory layouts for many internal data structures used in components such as DOM, CSS, layout, and painting. Google says that Blink, the rendering engine in Chromium, now "avoids a lot of useless churn" so as to make better use of the CPU caches. In the areas where memory handling previously relied on garbage collection in Oilpan, like the DOM (document object model), the team has expanded that by shifting from using malloc (memory allocation function) to Oilpan entirely. For those wondering, Olipan is the garbage collector in Blink. Some of the memory management and caching improvements Google made are fundamental to good code optimization. If you recall, recently, a senior Microsoft engineer also pointed out many of these issues in apps that slow Windows down. There are also improvements in handling strings within the renderer; the hashing method was updated to rapidhash, which is said to improve performance. For when rendering tasks become inherently expensive, such as computing CSS styles for various elements, Google adds that caching techniques have been enhanced to achieve higher cache hits and fewer misses.
    • Laptop users, this appears to be single-sided so it should fit even in cases with thin slots.
    • Apple wouldn't be what it is today without China either...
    • I am struggling with the game, after playing it again after so many years.... feels so out of space.
  • Recent Achievements

    • Week One Done
      Uranus_enjoyer earned a badge
      Week One Done
    • Week One Done
      jfam earned a badge
      Week One Done
    • First Post
      survivor303 earned a badge
      First Post
    • Week One Done
      CHUNWEI earned a badge
      Week One Done
    • One Year In
      survivor303 earned a badge
      One Year In
  • Popular Contributors

    1. 1
      +primortal
      428
    2. 2
      +FloatingFatMan
      195
    3. 3
      snowy owl
      191
    4. 4
      ATLien_0
      184
    5. 5
      Xenon
      141
  • Tell a friend

    Love Neowin? Tell a friend!