CWAP book comments after first read
Last Post: January 7, 2005:
After reading the CWAP book I have some comments to make on the content of the text. These comments are for constructive purposes to be included into an errata or for the next revision.
First, I thought the book was very good, easy to read and digest but a lot of material at the same time to digest. I have learned a few new things from this book, even outside of the 802.11 stuff which is good. The book will make an excellent reference book regardless if one takes the exam or not.
The text does seem to have an issue with flow of information and inconsistent use of acronyms or abbreviations.
If one is reading this book for the first time and this is the only book on 802.11 MAC protocol he/she has ever read it can be confusing if not frustrating at times in my opinion.
Here are some examples:
Page 51 TU are defined here but are Ã¢Â€ÂœformallyÃ¢Â€Â defined on page 174. Is this for reinforcement?
Abbreviations such as MPDU and MSDU are sprinkled in earlier chapters but not Ã¢Â€ÂœformallyÃ¢Â€Â defined until pages 97. and MMPDU again in chapter 6 pg.160
And not Ã¢Â€Âœreally in detailed definedÃ¢Â€Â until chapter 5 pg 135 under Nomenclature.
Since I read the Cisco Press 802.11 book these terms mentioned in earlier chapters like chapter 3 pg. 78. are not foreign to me but to the new reader these terms used in abbreviations without the name or definition early on can be confusing.
Pg. 217 refers to SFD but its meaning is not Ã¢Â€ÂœformallyÃ¢Â€Â defined until page 223.
BSA used in chapter 4 pages 104 and 106 but finally defined at the end of chapter 7 pg.210 in the Key Terms.
The Key terms from Chapter 4 Ã¢Â€ÂœConfiguration Options and Protection MechanismÃ¢Â€Â
Appear to be the key terms from one of the earlier chapters or are these more of Ã¢Â€Âœplaces to define some of the abbreviations???Ã¢Â€Â
DSSS-OFDM PDU type DSS-OFDM PPDU finally defined on page 232. Even though they are discussed in preceding pages.
Key Terms such as QAM and FSK at the end of chapter 8 are listed but not defined or discussed in chapter text nor in index and as for FSK did you mean Phase Shift Key or Frequency?
Pg. 306 Figure 10.7 Wrong figure heading??? Uses Laptop Analyzer Scenario again but for the WIDS figure???
Page 354. PAC finally defined. I am sure I saw this earlier in the book.
I do have some debate about the LLC usage. One question is in your book on page 372 you have LLC dated in 1998. I recall that LLC is much older than that. Are you referring to an update to the LLC standards?
From my experience in the wired world LLC Type 2 is used for netbios/netbuie or pure SNA type communication. As for seeing it on a wireless LAN chances are it is unlikely for the SNA stuff but maybe the analyst will see some for NetBIOS. LLC type 3 I have never seen used. I believe LLC 1 is the predominate for wireless. I am getting from your text that LLC is mainly used as a protocol discriminator(which is what SNAP was created for) between the 802.11 and 802.3 worlds and that should be the focus. Not on LLC 2 or 3 sequencing. I understand you want the readers to be well rounded but LLC 2 is a chapter in and of itself with mod 128 sequencing and REJ, Poll final, RR frames supervisory frames etcÃ¢Â€Â¦ ..
Also, the mentioning of using LLC for acknowledging of frames is used. That brought up one thought. Why bother with LLC 2 for frame ack on certain apps if at the LLC sub layer if the 802.11 MAC layer ack for every frame anyway?
Actually there is some room for debate and open discussion on TCP over wireless LANs. Remember TCP was originally designed for slow, circuit switched, x25, 56k and the like type of dirty circuits. TCP has flow control(slow start et. al.) to compensate for slow, dirty congested links, just like a wireless link. Actually in the world of fast and gigabit Ethernet TCP is more of a hindrance because it was designed for congested links and hence eats up all the bandwidth the stack can allocate. That is why such enhancements like ECF bits in the IP header for Explicit Congestion Notification in TCP and other slow start enhancements plus development of HS TCP or ETCP make running TCP on a wireless LAN much more efficient. So stating in the text
Ã¢Â€Âœthe problem with client/server connectivity is that it operates over TCP/IP, which is cumbersome over a wireless networks. TCP uses relatively large header and adds significant delays when reestablishing connections while the user traverses coverage holes..Ã¢Â€Â
This statement is somewhat subjective. There is Van Jacobson header compression in use(at stations and routers) today and again with the enhancements to stacks and TCP in general the ability to restart a connection is very fast.
My final vent which is not directed to the CWNP but to the general flow of packet detail in most books. Most of the books always take the OSI top down approach is detailing the PDUs and layering in operation. For most protocols from layer 7 to 3 this is okay. However, after reading many many protocol books over the years I find that for the lower level layers 1 and 2 working from the bottom then up is better. I would rather learn the signaling and coding first then follow that up through the MAC layer. It just seems simpler for me to lean this way for the more difficult PHY and MAC processes.
The cisco press book and the CWAP follow the rest of the industry where they discuss the MAC and all its fields first then later on go into the nuts and bolts of the physical layer attributes. For example the discussion on Ã¢Â€ÂœprotectionÃ¢Â€Â seemed a little confusing for first it is discussed at the MAC layer with RTS/CTS and CTS to Self then in the various PHY readings(from both books) there is inherent protection defined just by the presence of a certain modulation or service field or rate specification..
All in all I find the CWAP guide a very good read and as an all in one study guide this is refreshing.
When I did the CNX exams, Ethernet and Token-Ring, there were no study guides but a recommended reading list of like 15 books. Plus you not only had to know the MAC but LLC in and out and upper layer protocols like IPX and IP. Plus you had to know how to recognize these frames, layers 2-7, in a capture decode in hex only format and know the offset of where the LLC portion or IP portion of the frame is and what a corrupted frame looked liked in hex. Plus, do hex to binary to decimal conversions. Ugggg. The difference here is that you have to know all the RF PHY stuff in more detail than Manchester and Manchester differential. Personally, maybe the CWAP exam should have been JUST the MAC stuff:)))
So, the CWAP book is a great all in one study guide and reference and coupled with the Cisco press book you should be good to go.
BTW are there any trace file decodes in the exam? I remember asking this but forgot..
I hope I pass this examÃ¢Â€Â¦ This is defiantly not easy..
Many thanks for your THOROUGH and constructive feedback - it is greatly appreciated. We love this kind of feedback.
I made a written list as I went through so that I wouldn't forget anything....and hopefully I won't.
1. Acronym definitions -
Here are our top excuses for this problem:
a. We tend to work on chapters out of order
b. We have multiple authors and editors
c. The material is complex and there's nobody technically astute enough on the topic to review the book en-total when we're done
d. A book of this depth takes a very long time to write and the authors get exceedingly tired of looking at the material
I know these are just excuses, but they're the truth. We will try to clean that stuff up in an errata sheet. thanks a million for pointing out those detailed places to save us mucho time.
2. Our focus of putting LLC info into the book was as a "fill in the mental gaps", not as a thorough lesson in the LLC sublayer. Since Criss Hyde, our technical editor, contributed greatly (authored most of it) to this section, I will yield the floor to him...mainly because if I don't and I mis-speak, he'll whop me over the head...and I get enough whopping over the head from Kevin already.
3. On the ordering of the chapters.....plain and simple - I got outvoted. I wanted it to start at L1 and go up the stack. I was read the riot act over that one. Nevertheless, I lost...and you can see the result for yourself. :-) Sorry, I tried.
4. Decodes on the exam....
You bet. The exam has many decodes from many vendors. The one thing I should mention here is that the things we showed on the exam COULD have been shown easily on ANY anaylyzer that does 802.11 frame decoding, so the aesthetics of seeing an analyzer that you're not familiar with shouldn't throw you at all.
Again, I want to offer my thanks for great feedback - this is the kind of feedback that makes the content better.
No probs. It is a LOT of material to cover in one text. I can understand a lot of material on one text/subject like a book on eigrp or ospf but you cover many areas/subjects as well.
as for reason D - I understand about it being tiresome, even today how can I be reading about LLC.. again... but I understand especially on the phy stuff. uggg.
See us packet heads do think alike, sorry you got outvoted on the chapter sequencing.
It is a lot to swallow so I will probably read it two or three times, throw in another cisco press read and lab things with the analyzer during the same time before attempting the exam.
PS did you get my email and private message about the cert board?
Thanks again on a great book and look forward to others.
didn't get the PM or email....please resend PM to this forum.
I agree that the LLC text that made it into the book is overly general. The next edition should focus on bridging LLC as discussed in IEEE 802.1H and RFC 1042.
I insisted that LLC be included in the CWAP study guide because LLC is a big problem for bridging between IEEE 802.3 and 802.11 and not all vendors handle it the same way. As protocol analysts we see LLC taking up space in 802.11 frame bodies and in some but not all 802.3 frame bodies, even for the very same "frame" before and after bridging. We should be able to account for this and explain it.
It is easy for us to think that properly bridging layer three routed protocols such as Appletalk should be nothing to advertise. Thus a wireless router that includes "Appletalk support" must route Appletalk. A prominent WLAN commentator recently published just this mistake. For Asante's explanation of their wireless router "Appletalk support" please see this forum thread:
Your comments are much appreciated. Keep them coming.
Pg. 32 under the RTS/CTS paragrapgh. Last sentence.
"RTS/CTS may also used"
could be RTS/CTS may also be used
Pg. 56 under Null Function frame: first sentence
"coordinator that it has not data to transmit."
could be "has no data to transmit"
Pg. 77 Figure 3.10 WEP Data Frame Body structure Using WEP
The frame byte counts in the illustraton has the IV of 4 even though it says 3 below plus the key ID the paragraph on pg 76 references the IV to b 4 octets. This can be confusing for most IV discussions on WEP always state 24 bits 3 bytes and don't mentiong the key id byte.
The same for pg.79 figure 3.11 where the frame shows IV 4. Again some confusion could arise when using the term IV for most folks thing 24 bits(3 bytes) and not the key code byte.
Page 80 802.1X/EAP section first paragraph last section refers to figure 3.2(open system trace) did you mean figure 3.4(actual 802.1/eap trace)??
Good look on the book jsicuran. This is tough stuff. My hats off to those that pass this cert. I am back to studying was going to do some Cisco certs but decided to stick with this.
Got a question on pages 222 and 223, the two figures (8.4 and 8.5) are the same? Is this correct?
Aren't long preambles synch fields 128 bits with Scrambled 1's?
This fist edition remindes me of some of the cisco ccnp/ccdp books from cisco press circa 2000 error wise. I mena man cisco press these guys are big and their books had a lot of errors.
I know it is a lot of specific and broad info for a focused study guide so it is difficult to get everything perfect. All I can say is that if the quality is on par as that of the cwna book then this book is going to be a winner. I only hope what is in there is on the exam. Protocol analysis in general and in and of itself is a very subject art and science. Meaning there are many ways to get to a conclusion or many different symptoms can equal a cause. Trying to test that is very difficult. For example just on protection or power managment a test question can just show a packet and then you have to decide what type of protection or power management is used. Or vice versa.
If the exam was just on frame fields and state machines then it would be a little easier to focus.
There can be an entire exam on just protocol analysis techniques, which is not really discussed in the book. Like when and where to place an analyzer. Using two analyzers for each side of an AP or segment. When and how to descern when the problem is not an layer 1 or 2 but 3 or seven problem. How to discriminate symptoms and problems.
I am curious how such things are going to be tested in the cwne hands on lab exam. I did the ccie once and it is subjective and a lot of matrial to know but you know you are going in to configure routers and switches. Your configs are graded against a very specific script and the exam is designed for only one right answer per topic. How do you do that with an RF based exam?
Man I just hope I pass this cwap thing. Good luck.
Yesyou are right this CWAP cert is a good thing to have The book is a good read despite the errors.
The way you are going you will pass it.
I was studying late last night and back at it this morning. This is fascinating. I am now really understanding how the APs and other wireless devices talk to each other and make the spectral wonder that is the talk of coffee shops world wide. I have a deeper appreciation for wireless (802.11)and the CWNP program because of it. I think it only adds more value to the CWNA/CWSP and the folks at Planet 3 can really be proud knowing that they are going about this the right way. QoS!
Happy Thankgiving, got my CWAP book and the turkey going at the same time!
cwap and turkey.. take a break or at least run the microwave for the turkey and get that RF going big time in the house:)))))))