Subject: tcp/ip

I'm rather late with my TCP/IP comments. The only points I have to make

come in the way of caveats because all other comments are quite accurate regarding the content of the TCP/IP v.4 exam.

One point I have, and this was my mistake: I thought from everyone's

postings that I could expect a collection of scenario questions from a pool of several different questions. Well, this was my misunderstanding because I was questioned on every scenario mentioned on this site. This did not end up being a problem as I passed the exam but, I was surprised while writing the exam and one great thing about this site is it's suppose to prevent surprises while you are writing the exam. My mistake and I can heave a sigh now that it's over.

 I had several name resolution questions, many on utilities but nothing

on their switches. No ANDING questions. I did not write down the subnet tables as I find it easier and faster to work from the calculator but, that's my preference. Besides, I spent this time jotting down everything else I could think of! There were three questions on identifying the number of subnets and host addresses from a subnet mask and the address classes were all class B and class C addresses. No special address classes.

Three or four TCP/IP printing questions e.g. lpr and lpq and daemon type

questions.

I took the entire time but had plenty of time left over to review the questions I had marked.

 I 've gone back over this dump site looking for references from other

people to the following question. The reason I came back this question in particular is I didn't pay it much attention until I came upon it on my own exam and, unfortunately as its been some weeks since I wrote, I'm sure I don't remember all the subtleties. Perhaps some other contributors can provide better detail but, these are the two uploads I cut from other's contributor's comments along with what I can remember of it myself.

 The other odd one involved four NT servers on 3 subnets. It asked what command had to be run on the two middle servers to configure their static routing tables. Each answer involved two commands, one for each router. Answers 1), 3) and 4) were adding network IDs for the routers local subnets so obviously wouldn't achieve routing across all subnets.

 Answer 2) was fine except for the subnet mask [or was it the gateway?]

for router 3, which had an incorrect value in one of the octets. I chose 2) because it was the best one, although none were exactly correct so far as I could see.

 I wasn't sure of the answer to the "route add" question as I couldn't remember the command parameters (I know now that it's route add {destination address} mask {netmask} {gateway}). However, I figured it out by looking at the answers and you could see for each of the parameters the correct solution showed two of the correct items across each solution. For example, in the following items, you can see there are 2 occurrences of Bin the first parameter, two occurrences of A in the second parameter and 2  occurrences of C in the third parameter.

route add A mask A B

route add B mask A C <-- correct

route add B mask C A

route add C mask B C

I will add: you are provided with a diagram of the network. The question as I read it didn't say what the complete objective of the static routing table entries should be but my assumption was: "What entries would you include so all subnets could communicate the entire breadth of the internetwork".

It looks something like this, based on my memory of the question:

subnet.......................router...................subnet....................router.......................subnet

  | |

| | |

address: |

address: | address:

www.xxx.yyy.zzz | www.xxx.yyy.zzz

| www.xxx.yyy.zzz

address: address:

www.xxx.yyy.zzz

www.xxx.yyy.zzz

The thing that threw me aside from the fact that none of the available routing entry answers were the same as the ones I had worked out on my note paper was that the answers that made the most sense to me contained the subnet mask addresses that I expected to be the gateway addresses. And the final octet in the routing entries for the mask (netmask) from the available answers ended in zeroes. Such as:

 route add (destination address) mask (netmask) (gateway)

route add(www.xxx.yyy.zzz) mask(www.xxx.yyy.000) www.xxx.yyy.zzz

 I hope this makes sense, I know its a little confusing at first glance. My point regarding the answer is that a subnet mask address that ends in zeroes i.e.www.xxx.yyy.000 denotes an routing entry for

THIS NETWORK ONLY or, in other words, a local network. So, when I saw it I wasn't completely thrown off.

 The reason I've gone into this much detail for this one question is the question posed and the possible answers provided on the exam are unlike anything from my course materials, Transcender, a PEP test or the Exam Cram book and I'm ticked at whoever writes these exams for Microsoft for throwing these curve questions at us from left field when all the study and preparation possible wouldn't help us when we see it on the actual exam.

 Anyway, the amusing thing about all this is I still don't know if I answered it correctly but, it did make sense to me. I would suggest that if you have the time to research this question further you might look into it in a bit of detail or, if you are taking the curriculum in classes rather than self study you might ask the instructor or maybe resource TechNet.

When someone else encounters this question on their exam perhaps they will remember some of its detail better than I and smooth out the rough edges for future exam takers.

 Hope this helps.

 Dave.