Network Warrior 228
Fatty writes "Entry level certifications such as the Cisco Certified Network Associate (CCNA) have become the source of many jokes to people in the industry, largely because of the seemingly inept people that proudly display their certifications. This is made worse by the volume of books geared only to get people through the exam. Network Warrior bills itself as the exact opposite — if the subtitle is to be believed it contains "Everything You Need to Know That Wasn't on the CCNA Exam". With everything from the architecture of the 6500 to layers 8 and 9 of the OSI model (politics and money), it does a pretty good job." Read below for the rest of Sean's views on this book.
| Network Warrior: Everything You Need to Know That Wasn't on the CCNA Exam | |
| author | Gary A. Donahue |
| pages | 598 |
| publisher | O'Reilly |
| rating | 9 |
| reviewer | Sean Walberg |
| ISBN | 9780596101510 |
| summary | A practical look at what you really need to know to run a Cisco network |
The CCNA exam is supposed to test a candidate's understanding of networking fundamentals. Over the years it has expanded to include more advanced material, and now covers networking theory, switching (including spanning tree and VLANs), and some of the intermediate routing protocols such as EIGRP and OSPF. Despite the breadth of content the exam doesn't (and can't) cover things that many network folk take for granted, even things like what the "demarc" is (short for demarcation point, the the place where the carrier's responsibility ends and yours begins). While the exam's topic list is broad, the level of detail is shallow in most places. Someone may study spanning tree enough for the exam, but have no clue where to place their root bridge when they get into the real world.
It is for this reason that I found Network Warrior to be helpful. It's goal is to point out both the technical areas in which the CCNA falls short, and to teach the reader the non-Cisco aspects of running a network.
Technically I found this book quite sound. There were a few things one might disagree with but nothing that detracted from the rest of the book. In several spots the author was keen to point out behaviors that deviated from the documents, such as in Quality of Service (QoS) and in upgrading certain modules in the 6500 chassis. He also illustrated where the theoretical concepts on network design fall short in the real world.
Routing and switching takes up the first third of the book. The switching section is largely a review of the CCNA material with some notable exceptions. First and foremost is a chapter exclusively on autonegotiation. The CCNA exam may only discuss how to set a port to a fixed speed, but anyone who has worked with a network for more than a few weeks will have run into a speed or duplex mismatch. This chapter explains some of the history behind Ethernet and its relevance to autonegotiation, explains how it works, how it fails, and how to recognize the problem, and finally offers advice on when and where to use autonegotiation.
The second major deviation from the CCNA switching syllabus is in depth coverage of Etherchannel and spanning tree (STP) Both of these protocols are integral parts of network design and operation, but the exam barely touches Etherchannel and doesn't get into the complexities of spanning tree (though this changes with each iteration of the exam.) Network Warrior provides techniques and a demonstration of finding a layer 2 loop. Surprisingly though, there is only mention of standard 802.1d legacy spanning tree and some Cisco extensions such as Per VLAN STP and backbone fast, and no mention of the newer standardized enhancements of 802.1s/w (rapid spanning tree and multiple spanning tree) which have been in common use and have been put on the latest version of the exam (released after this book went to press)
The third deviation is the inclusion of CatOS commands instead of just IOS like the exam. As the author repeatedly points out, CatOS is in use on many 6500 chassis and is still in active development, so there is no reason not to know it. This theme continues throughout the book whenever the 6500 is used as an example, which is often.
The routing chapters are full of new material. The sections on the routing protocols themselves are short and don't add much beyond what the CCNA certification teaches. Redistribution and route-maps, however, are well explained. These two technologies which can be used separately or together can be found on almost any network and are very complex. I thought these sections were well done, as they gave enough details to be practical without getting down into all the different scenarios. Tunnels make an appearance in these chapters, which themselves aren't very complex, but aren't a part of the CCNA blueprint.
At this point, roughly page 180 of 550, the rest of the material isn't found in the CCNA blueprint.
Part 3 of the book is all about multilayer switching, specifically the 3750 and 6500 platforms. In particular the description of the 6500 architecture is much more succinct that can be found by searching on Cisco.com. There is an in depth explanation of how the various backplanes on the chassis works, which leads to an explanation of how to determine which cards are slowing down your switch.
I think the hidden gem of the book is part 4, though, which is all about telecom. In these chapters are an explanation of how carriers operate and how to speak the lingo of telecom techs. Even though networks are moving to Ethernet based services, traditional DS1, DS3, ATM, and frame-relay networks are still commonplace. The book has a solid explanation of how TDM based circuits actually work, the various options available to you, and how to properly order and troubleshoot them. I think back to when I was getting started in this field, and dealing with carriers was difficult.
Quality of Service, the features that let you guarantee and limit bandwidth to different types of traffic, have a section in this book too. The book largely focuses on the simple weighted-fair queuing (WFQ) and the current class-based WFQ with low latency queuing for voice. Configuration instructions can be found on Cisco's site easily enough, but Network Warrior delves into some of the behavioral aspects the documents shy away from such as when the queuing mechanisms actually get used. There is also a solid look at how to make sure the QoS is working as intended.
In the middle of all of this are chapters on the firewall and load balancing modules for the 6500, the PIX firewall, and IOS based load balancing. For someone with an ecommerce slant these might prove helpful, but given that these topics are books in themselves, it's hard to do them justice in a few chapters.
The last part of the book is on network design, which encompasses not only the steps needed to build a network, but also planning IP address allocations and how to pitch your ideas to management. Again, the book is not trying to be the definitive text on the subject, but it manages to impart a few words of wisdom, especially the so-called "GAD's Maxims", and "How not to be a computer jerk".
Well thought out examples were plentiful, along with anecdotes from the author, usually showing the consequences of doing things wrong. The illustrations did a great job of conveying the point at hand. Even though I've been doing this stuff for a while I learned several time saving techniques that I've already been able to put to use.
This is a great book for people just getting into the industry, with their CCNA or without. It offers practical advice rather than dry textbook like explanations which is a welcome change. Even those with a few years of experience under their belt will be happy reading through Network Warrior.
Sean Walberg is a network engineer and author living in Winnipeg, Canada.
You can purchase Network Warrior from amazon.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.
Interview Questions (Score:5, Insightful)
To my surprise, every single one of their obscure, imaginary-world answers were straight from sample MCSE tests. And after 10 more years working in a mixed environment, those questions still don't apply.
Then count yourself lucky.. (Score:3, Insightful)
Re:Interview Questions (Score:5, Insightful)
I just got lucky on the obscure questions they asked, they actually picked something that it was possible to come across in the real world (like, what command do you use to change a NT server to NT workstation?)
However I could swear I lost ability when i got my MCSE; so much of the stuff they test for is Microsoft "truthiness" that it causes confusion when you come across similar circumstances in the real world; if you are working with or for people who are Microsoft trained, you have to find some way to spin the real solutions so that it doesn't violate MS canon law.
Never did get my CNE; that was my next step until I decided to retire instead (I couldn't get a job doing anything fun, due to age barrier, my lack of desire to be management & everyone thinking I wouldn't be happy taking a pay & power cut from my previous job).
Re:Interview Questions (Score:5, Insightful)
I've seen the point argued back and forth on Slashdot. The anti-cert people say that there's little value in a cert that can be crammed for, a cert that doesn't really certify that the holder knows what he's doing. There are plenty of people with fancy certs on the wall who don't know what they're doing, just like there's plenty of people with no certs who are shit hot at what they do. The pro-cert people say that the certs serve as a measuring stick for non-techs who are looking to hire techs, a way of making sure that a candidate has a minimum level of experience before putting them through a serious evaluation. There's also the arrogance of geeks who think they don't need to bone up on theory and there's nothing more dangerous than the problems caused by what they don't know they don't know. The pro-certs people argue that the process forces you into a structured method of learning the topic.
I'm hip-deep in the process right now and I'd say it's a mixed bag. I think that the classroom instruction is good since it gives you a conversational environment to work through problems instead of just hitting the books on your own. The instructor, if he has real world experience, can also give you pointers you'll not find in the book. The bad part of all this is the testing. You can read the entire book, do the sample questions, and still be blindsided by the real test. The questions themselves are more designed to trip you up on stuff you know than really test you to see what you know. The technicalities and bullshittery of these questions is as bad as the worst tests endured in college.
From the cynical side, I've been told that the real scoop behind the certs is that companies like Microsoft want to make them seem like they have value so they want a high fail rate. If someone gets one, they should feel like they sweat blood. Now you can either make an exam tough with fair and exacting questions or you can use cheap tricks to fuck people up. Microsoft seems to prefer cheap tricks. And what's the worst thing that happens when someone fails? They pay to take the test again.
Re:book smarts vs real world smarts (Score:2, Insightful)
Re:Then count yourself lucky.. (Score:3, Insightful)
Re:Interview Questions (Score:4, Insightful)
Re:What value DO the entry level certs have? (Score:3, Insightful)
Re:I never ever got the point of Cisco certs (Score:2, Insightful)
That is a very good point... (Score:3, Insightful)
Re: OSI? Do you mean TCP/IP? ;-) (Score:3, Insightful)
For example, in TCP/IP, there's only one layer above the transport layer, which makes OSI layers 5, 6, and 7 all the same. I suppose you could argue that layer 6 never really existed anyway, but that's neither here nor there.
Similarly while there's a technical distinction between layers 1 and 2, in common practice they are too interrelated to speak of as separable layers.
Re:What value DO the entry level certs have? (Score:3, Insightful)
Sorry but no you don't. You learn Cisco's way of doing everything which is painfully different from everyone else from Nortel, 3com, HP, Brocade, Adtran, and a number of other vendors.
For every solution there is a standards compliant way to do it and there is a Cisco way to do it. To Cisco's credit their products support the standards but one need only look at the removal of CDP from HP ProCurve products to see why the Cisco way isn't always the best way. Subnetting into VLANs and trunking is basically the same from vendor to vendor. The Cisco exams are more concerned about how you setup LACP on a Catalyst 6500 running IOS 12 vs an older Catalyst running IOS 10 with the biggest difference being syntactical. The old way for instance you had to wipe an ACL completely and add it all back when you wanted to make a change. The new way you just remove the ACL entry you don't want. A much saner way. That's just one example. Cisco isn't even consistent among their product lines. Command sets are different with new releases of IOS, sometimes making it more friendly but I would think consistency would be a better approach.
Is it any wonder my Proxim APs have a very similar command set to my ProCurve switches?
So I would say that knowing the commands to setup LACP doesn't necessarily teach you the concept behind why you would want to do that and get into how it affects your STP setup.
I will say that the CCNA is still one of the better certs out there. Certainly light years better than any of the Microsoft or Oracle certs.
Comment removed (Score:4, Insightful)
Re:Interview Questions (Score:3, Insightful)