CCNA Question.

[polm]

Which of the following are not fields in an IP packet ? (Choose three.)

A. Flags

B. Length

C. Protocol

D. Source Address

E. Header Checksum


Now I was thinking that maybe they were trying to point out some differences in Packet vs. Header... but I just cant figure this one out.

The 20 byte IP header contains ALL of those feilds. What would the "packet" fields be ?

http://www.erg.abdn.ac.uk/user...t-pages/ip-packet.html <---IP Header

 

[spidey07]

its a trick question.

The "packet" is everything above layer 3. And hence it doesn't include the IP header.

So the packet wouldn't contain protocol, source address or header checksum would be my guess. TCP header does contain length, flags and header checksum.

There's a whole bunch of OSI questions like this on the test. And quite frankly I think they are worded very tricky and downright misleading.

 

[polm]

According to the CiscoPress Book. The Answer:

A. Flags

C. Protocol

E. Header Checksum

I really don't think this question is valid.

 

[polm]

I think the question now becomes: What was the Author really trying to ask ?

 

[polm]

TCP Header : http://www.freesoft.org/CIE/Course/Section4/8.htm

Does NOT have : Flags, Length, Protocol, or Source Address. It DOES have a Checksum (combination of Header and Data)

UDP Header : http://www.erg.abdn.ac.uk/user...se/inet-pages/udp.html

Does NOT have : ALL of them.

 

[polm]

It can't be packet vs. header. Packet has a header.

http://en.wikipedia.org/wiki/Packet

Time to send the Author of this book ,Steve McQuerry, an email.

 

[spidey07]

Originally posted by: polm
According to the CiscoPress Book. The Answer:

A. Flags

C. Protocol

E. Header Checksum

I really don't think this question is valid.

Its valid, it was asking you if you understood what is an IP header and what is a packet. A IP header wraps the entire packet.

A packet doesn't have to be tcp or udp, those are just two different IP protocols of which there are many (icmp, ospf, multicast, eigrp, rip, gre, ipsec, rip, igrp, etc)

Tricky? yes. Valid - yes.

those links you provided are really incorrect. They are calling a packet a datagram and they are not the same.

The OSI questions are really tough. Back when I took the CCNA I scored 100% on all portions except OSI where I got like 50%.

 

[spidey07]

Originally posted by: polm
According to the CiscoPress Book. The Answer:

A. Flags

C. Protocol

E. Header Checksum

I really don't think this question is valid.

I think I understand where the question was coming from. While a packet (everything above layer3) CAN include those things, it isn't required as it depends entirely on what kind of packet it is.

 

[polm]

But how does "Header Checksum" apply ? I would agree more with your proposed answer , and chosen "Source Address" .

Your explanation makes good sence to me. The "Packet" is protocol independant.

This is directly from the book.

The term packet is used loosely to refer to a peice of data. However, this book consistently refers to data formed in the transport layer as a segment, data at the network layer as a datagram, and data at the link layer as a frame.

Source address , seems, at the IP level to be required by ALL protocols.

Then again, the word "Header" in the answer Header Checksum may immediately disquialify it. Seeing that the question seems to be an effort to seperate the necessary header bits from the protocol independant "packet" bits.

any thoughts ?

 

[spidey07]

absolutely source address is required by the network/IP layer (layer 3)

But you can have a TCP or UDP header as well and each header has its OWN checksum for the rest of the data that it is encapsulating.

Its confusing.

 

[polm]

Here is one more that I had trouble grasping. I understand the process of establishing a TCP connection using SYN and ACK.

But the book asked this question, and the answers really had me scratching my head: (the commas and semi-colons ARE directly from the book)

Select the statement that correctly orders the steps to establish a TCP connection,

A SYN=0, SYN=0, ACK=1; SYN-0 ACK=1

B SYN=1, SYN=1 ACK=0; SYN=1 ACK=0

C SYN=1, SYN=1 ACK=1, SYN=0 ACK=1

D SYN=0, SYN=1 ACK=1; SYN=0 ACK=1

E SYN=1, SYN=1 ACK=0; SYN=1 ACK=1

It is the commas and semicolons that really have me most confused.

 

[polm]

Dmanit..heres another one that I am confused on . All from the same chapter BTW.

What TCP characterisitics are not found in UDP ? (Choose three.)

A Connections

B Flow Contol

C Data Transfer

D Multiplexing

E Fragment-Free

F Realiable Transfer

Now I was SURE i knew this one (I chose A,B,F) ...I was wrong according to the boook.

 

[polm]

Originally posted by: polm
Here is one more that I had trouble grasping. I understand the process of establishing a TCP connection using SYN and ACK.

But the book asked this question, and the answers really had me scratching my head: (the commas and semi-colons ARE directly from the book)

Select the statement that correctly orders the steps to establish a TCP connection,

A SYN=0, SYN=0, ACK=1; SYN-0 ACK=1

B SYN=1, SYN=1 ACK=0; SYN=1 ACK=0

C SYN=1, SYN=1 ACK=1, SYN=0 ACK=1

D SYN=0, SYN=1 ACK=1; SYN=0 ACK=1

E SYN=1, SYN=1 ACK=0; SYN=1 ACK=1

It is the commas and semicolons that really have me most confused.

Ahaaa...

SYN does not necessarily mean Sequence Number.

So C is correct, because first the SYN is set , and the ACK isnt, then then receiving station sends it's ISN with a SYN so the bit is set, and it also sends an ACK back. The original station sends and ACK but doesnt send a SYN
cuz "session establishment" is done.

Right ?

I still don't know what the semi-colons are supposed to mean...but I guess it doesn't matter.

 

[TechnoPro]

Originally posted by: polm
Dmanit..heres another one that I am confused on . All from the same chapter BTW.

What TCP characterisitics are not found in UDP ? (Choose three.)

A Connections

B Flow Contol

C Data Transfer

D Multiplexing

E Fragment-Free

F Realiable Transfer

Now I was SURE i knew this one (I chose A,B,F) ...I was wrong according to the boook.

Answer, please?

I'd agree wholeheartedly with A and F. But the third one is eluding me.

 

[spidey07]

Originally posted by: TechnoPro

Originally posted by: polm
Dmanit..heres another one that I am confused on . All from the same chapter BTW.

What TCP characterisitics are not found in UDP ? (Choose three.)

A Connections

B Flow Contol

C Data Transfer

D Multiplexing

E Fragment-Free

F Realiable Transfer

Now I was SURE i knew this one (I chose A,B,F) ...I was wrong according to the boook.

Answer, please?

I'd agree wholeheartedly with A and F. But the third one is eluding me.

I'd guess B, E, F

UDP has a connection (even though its called connectionless)
TCP has flow control thru windowing
Both can do data transfer
multiplexing has nothing to do with it
TCP can demand no framgement (I believe, I think that is a function of the IP layer/flags)
TCP is reliable

 

[TechnoPro]

Originally posted by: spidey07

Originally posted by: TechnoPro

Originally posted by: polm
Dmanit..heres another one that I am confused on . All from the same chapter BTW.

What TCP characterisitics are not found in UDP ? (Choose three.)

A Connections

B Flow Contol

C Data Transfer

D Multiplexing

E Fragment-Free

F Realiable Transfer

Now I was SURE i knew this one (I chose A,B,F) ...I was wrong according to the boook.

Answer, please?

I'd agree wholeheartedly with A and F. But the third one is eluding me.

I'd guess B, E, F

UDP has a connection (even though its called connectionless)
TCP has flow control thru windowing
Both can do data transfer
multiplexing has nothing to do with it
TCP can demand no framgement (I believe, I think that is a function of the IP layer/flags)
TCP is reliable

Please explain the bit about UDP having a connection but is called connectionless? Not that I doubt your expertise, but I am truly confused.

 

[spidey07]

UDP has acknowledgements...meaning server sends data and client acknowledges data, but it relies on the upperlayer protocol for sequencing.

TCP as a TRANSPORT layer does the sequencing, acks and sliding window, retransmissions for data transfer - the upper layer protocol doesn't need to worry about it, tcp handles it as the transport layer.

So UDP does indeed form a connection on a transport layer - there are PORT numbers involved that when combined with IP addresses create a connection. But it is called CONNECTIONLESS because as a TRANSPORT layer it is not responsible for the retransmission of data.

confusing?

TCP handles the actually transfer of data via sequencing, acknowledgements and sliding windows.

UDP does not, it merely ACKS data.

And that is the key between connection oriented and connection-less oriented transport layers. Most of this stuff becomes stunningly clear when you look at sniffer traces of the various kinds of packets and really see what is going on the wire and get to see each of the OSI layers broken out.

There is a reason that this OSI stuff is ground into network people over and over and over again. Its because it is so dang important and is the very foundation that network communications is built on. Kinda like if you want to be a bridge builder you better know the basic properties of the materials involved and the math as well.

 

[polm]

Sorry for taking so long with the book's answer:

Answer:

A Connections

B Flow Control

E Fragment-Free

 

[TechnoPro]

I'm throwing in the towel on this one... The answer would suggest that UDP is a reliable means of transfer. This goes against much of what I was taught.

 

[TheCause]

Every book that Cisco produces is filled with stupid flaws. What kind of group decides that ping is suddenly an acronym anyway!!! AGGGHGHHH! *sigh. There is a book that is publicly available for download that covers Cisco materials. It's a project totaly independant of Cisco Systems.

 

[spidey07]

Originally posted by: TheCause
Every book that Cisco produces is filled with stupid flaws. What kind of group decides that ping is suddenly an acronym anyway!!! AGGGHGHHH! *sigh. There is a book that is publicly available for download that covers Cisco materials. It's a project totaly independant of Cisco Systems.

heh, PING has always been an acronym in network world.

 

[spidey07]

Originally posted by: TechnoPro
I'm throwing in the towel on this one... The answer would suggest that UDP is a reliable means of transfer. This goes against much of what I was taught.

UDP is not reliable from a transport layer, it relies on the upper layer protocols to do the retransmission.

Case in point - TFTP uses UDP. You can drop packets all over the place and the file is still transferred.

 

[TheCause]

Originally posted by: spidey07

Originally posted by: TheCause
Every book that Cisco produces is filled with stupid flaws. What kind of group decides that ping is suddenly an acronym anyway!!! AGGGHGHHH! *sigh. There is a book that is publicly available for download that covers Cisco materials. It's a project totaly independant of Cisco Systems.

heh, PING has always been an acronym in network world.

Not according to the original creator of the original ping program. Those who control the present control the past. Those who control the past control the future.

 

[spidey07]

???? What?

PING comes from submarines and we turned it into an internet standard for packet internet groper. It stemmed from some of the first RFCs describing internet control message protocols (ICMP)

 

[TechnoPro]

Originally posted by: spidey07

Originally posted by: TechnoPro
I'm throwing in the towel on this one... The answer would suggest that UDP is a reliable means of transfer. This goes against much of what I was taught.

UDP is not reliable from a transport layer, it relies on the upper layer protocols to do the retransmission.

Case in point - TFTP uses UDP. You can drop packets all over the place and the file is still transferred.

Your clarification makes great sense, thank you.

Would you happen to know of any good texts that go into the OSI model at both theoretical and practical levels?