A: Parsing Layer and Header Terminology
I posted a question about headers and TCP/IP layers a few days back. Here are the answers! Post with any related clarifications or comments.
Answers: A, D
The TCP/IP model exists in two versions. One model uses four layers, with the lowest layer called Network Access or Network Interface. The alternate model uses the same upper layers, substituting two layers – the Data Link and Physical layers – for the Network Access layer.
The Ethernet header is generally associated with the data link layer, which matches the upper half of the network access layer. So, the terms Data Link and Network Access both apply when discussing Ethernet headers. Those facts rule in the two correct answers, and rule out the one incorrect answer that mentions layers.
The PDU acronym, Protocol Data Unit, refers to any header/trailer and its encapsulated data. That encapsulated data may include headers from other layers. For example, a layer 2 PDU (L2PDU) includes the layer 2 header and trailer, plus the encapsulated data. That encapsulated data, and often does, include a layer 3 header, layer 4 header, and so on.
However, a LxPDU does not include headers and trailers from lower layer headers. So, a layer 3 PDU (L3PDU) would include the layer 3 header, layer 4 header, and so on, but not the layer 2 header.
The term frame, listed in the question, refers to a layer 2 PDU. An Ethernet frame is therefore a layer 2 PDU that uses Ethernet framing, namely an Ethernet header and trailer.
All these facts about PDUs combine to rule out the other two incorrect answers. The Ethernet header, part of a Layer 2 PDU, is not part of any Layer 3 PDU. However, the Ethernet header in a layer 2 PDU is not the only header in the layer 2 PDU.
I see why A is listed as a correct answer, and why the others are incorrect. However, I think the terminology for D is a bit misleading. In looking at your Cert Guide for ICND1, it shows the four layer TCP/IP “Network Access” layer listed as the ‘Link’ layer, and on the five layer TCP/IP it has Data Link and Physical as the lowest two layers splitting the ‘Link’ layer up. Now, I haven’t read the entire book yet, so it may be termed the Network Access layer in another chapter.
As for my response in the last post saying E could be correct, my thinking on that was that you didn’t count the other layers’ headings in the PDU since it was encapsulated as data the lower it goes in the OSI model. I stand corrected on that now.
Hey Matthew,
Thanks for the explanation. And in the same spirit of your comments… I do take a few more liberties with wording on things posted here in the blog, because there’s a way to talk it through. For questions that are printed in the book, or are in the practice tests that come with the book, I’m a lot pickier – although there will always be room for interpretation.
End of it all, if you have an accurate mental picture of which headers follow which, and which headers are/aren’t in a frame/packet/segment aka L2 PDU/L3PDU/L4PDU, then you’re probably in great shape. Which it sounds like you do.
Wendell
I think you just made up the term “Network access”, i do not see it on the book (100-105)
Hi Leo,
Sorry to hear that.
It’s on page 21 of the book, by the way.
Wendell
That was a great question and explanation, thank you! It just got clearer to me now.
Regards!
Hi Wendell,
Is “E” incorrect because Ethernet is not the only potential Layer 2 technology?
Hi Tim,
I’d count E as incorrect because:
“Layer 2 PDU” refers to the bytes that begin with the layer 2 header, including the encapsulated data.
The encapsulated data includes the IP header, TCP header, and maybe HTTP header in this case (given that it’s web traffic).
So, answer E, as worded, says that it includes only the layer 2 header, whereas in reality, it includes the layer 2 header as well as other headers.
Wendell
Never better time to make CCENT/CCNA exam(s) performance- based than now. All this tricky-pony wording will be gone once the exam becomes performance based. In the age of cloud- everything and work done on remote more than ever before we are still stuck with multiple choice questions and trickster wording?! The reason why D is the right answer is one more reason to make these exams performance-based.
The new exam(200-301) is out and it is not just performance based. My old technical manager used to say ” If your theory is strong, your practice will be.” I don’t see a way for performance-based only testing at CCNA level. I feel it will always be both based on testing conceptual understanding and performance. Just my 2 cents!
that was a good question, however, it did not specify to pick two correct answers. I’ve seen questions like this on an exam before and usually they want you to pick the best answer. Only if it specifies to pick two correct answers would you pick A and D.
I’m still confused. From page 21, it’s the Link layer in the original TCP/IP model that becomes the Data Link and Physical layers in the TCP/IP Updated model. The Network layer matches with the original Internet layer which is above the Data Link and Physical layers. Therefore, I’m not sure how D can be correct. Can you tell me what I’m missing? Thank you for your help!
Hi Michelle,
Well, I’m not surprised. I’ve been around a while, and over time, Cisco’s docs and courses have varied with what terms they prefer for that original bottom layer that matches the OSI physical + Ethernet. Anyway, look at that same page 21 in the book, to the gray highlighted text, and you’ll see the note that gives the two other traditional names for “link layer”: “network interface” and “network access”. Those have been used more prominently in the past, only mentioned in the book, but I happened to list them in this practice question.
Wendell
Hi,
If HTTP protocol is stateless, how the next responses messages of the Web Server do not have HTTP headers?
Reference: HTTP Protocol Mechanisms. Official Cert Guide CCENT/CCNA ICND1 100-105. Page 23
Thanks in advance.
Fernando N.
Hi Fernando,
I don’t think HTTP is stateless… does that clear up the issue?
Wendell
So E is a incorrect because a L2PDU have other headers from upper layers inside?
Yes, you are correct!
When people reference page numbers for the textbook. They are completely different from what is on the corresponding pages in the textbook that I have .example:You mentioned that “network access” was brought up on page 21. But in my book page 21 covers TCP Error Recovery and there is absolutely mention of even the word “network access”
I am using the 2020 version, have there been some changes? I’m a bit lost here
Hi Nicholas,
Yeah, in a perfect world, I’d have enough budget to update all 600+ blog posts from over the last decade every time a book revised. Short version: Whatever page number reference you saw was probably for an earlier edition of the books. You should see the publication date of this post just under the title, from August, 2015. We’ve released two later editions of the CCNA cert guide books (2-book set) since this blog post was made. So while the concepts here might be of use, the page numbers of course are old.
For the current books, look in the CCNA 200-301 Vol 1 book, pages 25-30, for the similar content in the current books.