It's a network world

I advocate people take CCNA to get the networked foundation first. Then if servers interest them other than networking invest in MS/Novell/Linux/Unix certs to suit.

Looking at the posts on the MS vs Cisco boards there is a trend..

Cisco = 167254
MS = 95243

My numbers might be off as I took a stab at it anyway..seems interest in networking is growing..

Find more posts tagged with

Save $250 on 2025 certification boot camps from Infosec!

Book now with code EOY2025

Button

Comments

Roguetadhg

Turgon wrote: »

I advocate people take CCNA to get the networked foundation first. Then if servers interest them other than networking invest in MS/Novell/Linux/Unix certs to suit.

Looking at the posts on the MS vs Cisco boards there is a trend..

Cisco = 167254
MS = 95243

My numbers might be off as I took a stab at it anyway..seems interest in networking is growing..

Isn't Network+ supposed to be the opening for Networking?

Turgon

Roguetadhg wrote: »

Isn't Network+ supposed to be the opening for Networking?

Just thought I would compare server NOS with Networks

Roguetadhg wrote: »

There Layer 7 guys I know don't bother to care about the lower levels. More so they have a "It's broke, not my fault" attitude.

Networking professionals had to get beyond layer 4 in 2002. Cisco was good at layers 1 -3, not 4. Today we need to be good across all layers of the OSI model.

jmritenour wrote: »

Agree with you, except I'd scratch Novell of the list of server NOS to pursue. Whether we're talking Netware, SLES, or OES, I don't believe any Novell certs offers much of a ROI at this stage of the game.

That's just not the case. Novell is holding in enclaves and if you want to prosper there you better know something about it as you will be up against time served Novell people. Plus they have niche products like sentinal. Beware of such sweeping statements.

shodown

This makes too much since starting at the lower levels of the OSI model, and if you choose servers to then go into which ever brand. This usually falls on death ears. I was one that started with the servers then going into networking. But after learning the lower levels it all made since.

Roguetadhg

shodown wrote: »

This makes too much since starting at the lower levels of the OSI model, and if you choose servers to then go into which ever brand. This usually falls on death ears. I was one that started with the servers then going into networking. But after learning the lower levels it all made since.

There Layer 7 guys I know don't bother to care about the lower levels. More so they have a "It's broke, not my fault" attitude.

jmritenour

Agree with you, except I'd scratch Novell of the list of server NOS to pursue. Whether we're talking Netware, SLES, or OES, I don't believe any Novell certs offers much of a ROI at this stage of the game.

afcyung

I think a fundamental knowledge of Networking is important but one could make the same statement about basic knowledge of servers and their functions.

ptilsen

I've seen you say this a few times, and I have to respectfully disagree. "It's a networked world" is, frankly, a platitude. It's no more logical an argument than "become a certified electrician; all computers and networking equipment run off electricity" or "learn assembly; all systems run on assembly code".

The fact of the matter is that there are many, many, many entry and mid-level IT jobs that simply do not utilize even CCENT-level networking knowledge. I would further argue that most non-networking IT jobs don't need a CCNA-level of knowledge.

Now, let me play devil's advocate with myself and explain why your advice will still lead to success in non-networking IT careers. Yes, indeed, "it's a network world". Most helpdesk people, DSTs, server admins, systems engineers, etc. need to know how networks work. They need to be able to identify and differentiate network problems from computer hardware and software problems. They need to know the OSI and TCP/IP models. They need a high-level, but thorough understanding of how computers and networking equipment all work together to make the Internet and organization networked computers systems work. Many generalists need to be able to configure switches, routers, firewalls, and VPNs even if networking is not their full-time gig. Any server guy needs to understand NAT and how firewalls work. Anyone designing any kind of server system needs to understand what the network needs will be. All of these needs are 100% met by achieving CCNA. This is why I still advocate that the vast majority of server/systems guys consider CCNA, and that most actually get it at some point in their career.

However, we come full circle to why I disagree with the recommendation that CCNA is the starting point.

First, most of the above needs I describe are largely met with Net+ or CCENT. The difference between CCNA and CCENT is largely that you have a deeper knowledge of how to configure Cisco routers and switches for "advanced" features and protocols. This is not relevant to many sysadmins' jobs, and is completely irrelevant to a helpdesk or DST job. It can actually cause some heartache as a professional who studies for and passes CCNA may spend a lot of time becoming comfortable with IOS and then never get to use it in production.

Second, CCNA is an intensive, mid-level certification. There is a lot of knowledge required to pass CCNA, and an entry-level guy who is just learning about IT is going to struggle. A+ is a far more achievable exam for the inexperienced and is far more relevant for doing entry-level jobs and for obtaining entry-level jobs. Net+ is a good follow-up that fills in a lot of the networking gaps the entry-level IT professional lacks. It's also much more attainable than CCNA, typically requiring only a few weeks of study for the uninitiated, potentially a few days for the experienced.

Finally, and overlapping with both points, there is so much information in CCNA that just isn't relevant to the average generalist. To this day, I've yet to encounter a need to configure, observe, or even understand a routing protocol. I've never met another systems/server/generalist IT pro that has needed to understand or use routing protocols. Obviously understanding routing is important, but knowing how to configure OSPF or even understanding what kind of protocol it is is completely irrelevant to my job, my job prospects, and my foreseeable career. I can't fathom using that knowledge. Even if I were consulting to or working for much larger organizations than I am, the routing would be handled by the service provider(s). Furthermore, the frequency with which I and other systems/generalists work with Cisco equipment is non-existent. I can configure STP or port security or link aggregation or whatever I need, but I don't see a real need to memorize the syntax and functionality.

So, my real problem with the CCNA is that it so heavily geared towards networkers. The truth is, the industry lacks a good equivalent that can help a professional understand networking thoroughly without being forced to learn the details of implementation. The same is true on the networking side: Every CCNP & above should know the high-level of how many MS and Linux technologies work, but not know the implementation details. They should know how DFS replication, Remote Deskop Services, WSUS, etc. can be relevant to network design without having to learn how to implement. The problem between both career tracks (and really, between MS & Cisco) is the scope of the exams. A networking/Cisco professional doesn't need to know WDSUtil syntax anymore than I need to know OSPF syntax.

Anyway, in my experience as a successful generalist with a near-CCNA level knowledge who has mentored many other generalists/systems guys from helpdesk to server admin, CCNA is not a logical starting point either for career purposes or knowledge purposes for a non-networking. The correct order is more or less A+, Net+ or CCENT, vendor/OS certs (MCSA/E, MCTS/MCITP, RHCE, Linux+, etc.), followed by either CCNA or further specialization. Keep in mind, the market value of certifications varies regionally, and I understand that the UK and Australia do not value Comptia quite as much as the US. I do believe that most professionals should eventually take CCNA, but not right away. And that's why I plan on sitting for my CCNA sometime this year.

So, to conclude my rant, I respect your experience and your opinion, but my perspective and experience lead me to disagree with it. I do not believe the CCNA is a logical starting point for an IT professional not looking to work in the Cisco world, and I've yet to see any compelling evidence or argument that will sway my opinion in that matter.

jamesleecoleman

jmritenour wrote: »

Agree with you, except I'd scratch Novell of the list of server NOS to pursue. Whether we're talking Netware, SLES, or OES, I don't believe any Novell certs offers much of a ROI at this stage of the game.

Would they offer a good ROI if the person worked at a school system that uses Novel? The highschool and community college I went to both used Novel.

ptilsen

jamesleecoleman wrote: »

Would they offer a good ROI if the person worked at a school system that uses Novel? The highschool and community college I went to both used Novel.

Sure they would, but so would BES certs at an organization that has a huge BlackBerry infrastructure. That doesn't mean either is a smart line of technology to invest your time, money, and career into. If the employer is paying for it and you see yourself there for a while, that's one story. If you're just looking to specialize and get into something, going with the certs in high demand makes a lot more sense.

it_consultant

ptilsen wrote: »

I've seen you say this a few times, and I have to respectfully disagree.

....

So, to conclude my rant, I respect your experience and your opinion, but my perspective and experience lead me to disagree with it. I do not believe the CCNA is a logical starting point for an IT professional not looking to work in the Cisco world, and I've yet to see any compelling evidence or argument that will sway my opinion in that matter.

Every server guy really has to know a lot of networking in order to do his/her job effectively. Used to be that the opposite was not true for the network guys. I remember when I first got certified CCNA we were learning about ACLs and applying them in routers and switches. I asked how this would effect MS Exchange etc. I was told "we don't worry about that stuff". Hell, even MS tests will ask you questions on ports that need to be opened on firewalls and other network specific questions.

There is a subset of network professionals who don't have to care about our silly little applications; not that many that a beginning network engineer can ignore the Linux/MS folks.

Truly, I think network engineers started getting a lot more diversified when Cisco started being a voice player. Then they had to start thinking about servers for things like voice mail, phonetree logic, and the like.

jmritenour

Turgon wrote: »

That's just not the case. Novell is holding in enclaves and if you want to prosper there you better know something about it as you will be up against time served Novell people. Plus they have niche products like sentinal. Beware of such sweeping statements.

I am a "time served" Novell person, and I don't believe any current Novell certification has any value. Bear in mind, I spent 6 years working in a K12 system that was still clinging to Netware, and began to migrate to OES as I was leaving. I'm not saying the technology itself doesn't still have have it's uses, just that getting a cert in it isn't going to pay off much in the long run.

Just like when there was a frenzy to come up with people that knew Cobol before 2000 in order to re-write code to be Y2K compliant -great if you already knew the language, you could make a killing. But for someone about to graduate college in 1999? That time was better spent on languages that had a future past some bug fixes.

Using your own metric, searching on job boards, I was only able to find 10 Novell related positions within a 50 mile radius of me. And I'm in the DC metro area - we're not exactly hurting for IT jobs here.

Turgon

jmritenour wrote: »

I am a "time served" Novell person, and I don't believe any current Novell certification has any value. Bear in mind, I spent 6 years working in a K12 system that was still clinging to Netware, and began to migrate to OES as I was leaving. I'm not saying the technology itself doesn't still have have it's uses, just that getting a cert in it isn't going to pay off much in the long run.

Just like when there was a frenzy to come up with people that knew Cobol before 2000 in order to re-write code to be Y2K compliant -great if you already knew the language, you could make a killing. But for someone about to graduate college in 1999? That time was better spent on languages that had a future past some bug fixes.

Using your own metric, searching on job boards, I was only able to find 10 Novell related positions within a 50 mile radius of me. And I'm in the DC metro area - we're not exactly hurting for IT jobs here.

And Im a time served Novell person of 14 years standing and a CNE to boot, and while I agree, Novell has had it's day, if you find yourself in shop that is entrenched in Novell and there are still many out there, you need to know something about it. eDirectory, Zen, Groupwise the works.

instant000

afcyung wrote: »

I think a fundamental knowledge of Networking is important but one could make the same statement about basic knowledge of servers and their functions.

+1000!

I've had the pleasure throughout my career to work on both sides of this ... continental divide.

Literally, the networking guy will say it's something wrong with the server, and the server guy will always say its the network (mind you, neither side has done any troubleshooting yet to confirm anything?! ... it's just a reflex response.)

Also, I just don't see how you can accurately troubleshoot a multi-part, network-based application, if you don't have some sound idea about how all the pieces should fit together. I'm not saying every server admin needs a CCNA, nor am I saying that every network admin needs an MCITP/RHCE, but at least some knowledge should be there about how these things actually work, and you won't find it if you stick strictly with "your side" of the house.

The network is a symbiotic relationship. If there are no applications running across your infrastructure, the infrastructure is worthless. So, it makes sense to be really good at getting those applications from point A to point B, and if there are problems, be able to pin-point them so you can get the app moving again.

The beauty of OSI is that you can literally dilly-dally with your applications all day long, with no care or concern about how the network actually works underneath. This works really great, until you start running into scenarios where application response and behavior is affected by that network. When you demo apps in an isolated local test VLAN (running at minimum 1 Gbps), then you deploy them to production across a WAN (running at MOST T-1) and wonder why the performance is suffering, and start complaining to the network team, please face-palm yourself, mmmkay?

And network team, you're not off the hook, either, for implementing IP re-numbers and such for things that "shouldn't" be affected, but who knew joe schmoe's server had some hard-coded setting for that? Or ... one of my more favorite ones, the ACL that's cleaned up in March, but since the session was active, it stayed active, and until a power event in September, the session was still active through the ASA. Now, the user app is broken, and network team hasn't done any recent changes to immediately know why. The server people are crying, and the network people are playing honest (haven't investigated anything yet, just pointing fingers). Of course, the problem gets fixed eventually, but still ...

I feel that situations like this could be avoided, if there was a greater knowledge about what the other side had to deal with. Believe it or not, people, not every server admin is an idiot, and not every network admin is out to make sure your applications don't work.

instant000

Turgon wrote: »

I advocate people take CCNA to get the networked foundation first. Then if servers interest them other than networking invest in MS/Novell/Linux/Unix certs to suit.

Looking at the posts on the MS vs Cisco boards there is a trend..

Cisco = 167254
MS = 95243

My numbers might be off as I took a stab at it anyway..seems interest in networking is growing..

Oh yeah, had to rep this one, too, started a nice discussion

ChooseLife

Turgon, would you also agree that network engineers should know OS and systems principles too?

Roguetadhg

ChooseLife wrote: »

Turgon, would you also agree that network engineers should know OS and systems principles too?

There's more to a network than just the physicals. Obiviously. A trained monkey can figure this out.

I think it's best to know something, atleast know it's there and have 'some' experience with it, while focusing on whatever you like best.

I've run into the "It's not our problem" scenario a lot- while I troubleshooting as much as I could (Because I didn't want a finger pointing game where nothing got fixed as tends to happen). I'd rather go to someone saying "I've rebooted, shut down and turned on, tried to use my profile, tried x, y, z, affects one person, searched google for a solution exhausted sources, used x website to test..." In the least, it would help skip steps. While I may not know anything about their position or software - when end users come to me I tell them "I don't know, never used it, and I know you know more than myself". I don't get any training with anything, I learn the software from end-users.

jmritenour

Turgon wrote: »

And Im a time served Novell person of 14 years standing and a CNE to boot, and while I agree, Novell has had it's day, if you find yourself in shop that is entrenched in Novell and there are still many out there, you need to know something about it. eDirectory, Zen, Groupwise the works.

I don't disagree with that at all. What I've been saying all along, however, is that I wouldn't advise someone just getting into the game spend any time/money/effort earning Novell certifications. And I stand by that.

jamesleecoleman wrote: »

Would they offer a good ROI if the person worked at a school system that uses Novel? The highschool and community college I went to both used Novel.

Missed this one last night. I say no, and the reason for that being that school systems pay for crap.

I left a K12 - where I had a senior role just below our department head - that was a Novell shops (Netware/SLES/OES/Groupwise/Zenworks, et al) in January last year to go back to the corporate world, and increased my salary by 40%. Granted, I did not hold any current Novell certifications, but none of my certs I earned while I was there ever got me more than a pat on the back.

Of course that's all relative, I'm sure some school systems/colleges have bigger budgets for their IT staffing, but in my area, school system budgets are being cut every year, and IT departments tend to suffer the most. No way I'd go back to that.

networker050184

Ptilsen, I beleive you have only worked in VERY small environments if you've never had the need to run a routing protocol? Have you never worked in an environment with more than one router? The service provider is only going to work from your DMARC out. You are still going to need routing inside your own network.

I'd also say that if you deem the topics covered in CCNA "advanced" you really don't know much about the networking world to begin with. The CCNA covers the foundations of getting a small network up and running (well, the router and switch part anyway, it doesn't even begin on firewalls). This is something that many "generalists" are responsible for. Getting anything above a couple routers and switches running is going to take a bit more than what the CCNA covers.

it_consultant

networker050184 wrote: »

Ptilsen, I beleive you have only worked in VERY small environments if you've never had the need to run a routing protocol? Have you never worked in an environment with more than one router? The service provider is only going to work from your DMARC out. You are still going to need routing inside your own network.

I'd also say that if you deem the topics covered in CCNA "advanced" you really don't know much about the networking world to begin with. The CCNA covers the foundations of getting a small network up and running (well, the router and switch part anyway, it doesn't even begin on firewalls). This is something that many "generalists" are responsible for. Getting anything above a couple routers and switches running is going to take a bit more than what the CCNA covers.

I really wouldn't leap to this conclusion very quickly. I have worked in massive environments where the internal routing was QUITE simple when compared to an ISP style networker. I wouldn't call CCNA topics advanced as much as I would call them, in many cases, completely irrelevant - which adds confusion to the world of entry level IT workers.

CCNA is a must have because it is fairly rigorous which has garnered it a certain level of respect. The closest non-vender cert is Network + which is a comparable joke.

On a regular basis I get junior(ish) engineers who achieve CCNAs and off to the races with ACLs and VLANs we go. Until I tell them to sit down and shut up, give them the console to an HP and say "Here is a different kind of switch, I am going to teach you how to do everything you just learned more easily and clear up the Cisco cloud of confusion hanging over your head". Or my personal favorite, "this is spanning tree protocol and this is how we turn it off".

My biggest complaint about CCNA is that it packs a lot of good information without a higher overview of the application for a lot of that knowledge. The first question I ask newly minted CCNAs is "why?". Why do we need to do this or that? What benefit will it provide us etc and so forth.

networker050184

it_consultant wrote: »

I really wouldn't leap to this conclusion very quickly. I have worked in massive environments where the internal routing was QUITE simple when compared to an ISP style networker. I wouldn't call CCNA topics advanced as much as I would call them, in many cases, completely irrelevant - which adds confusion to the world of entry level IT workers.

Of course there are plenty of simple networks out there, but even with only two routers with some subnets behind them you need a routing protocol. I guess you could go with static routes, but that's not a very good idea. The CCNA teaches you how to set a simple network like this up.

And I won't even get started on turning off STP.

What I've learned from my time working in IT and being on this site is the less you know about a network the more simple you think it is.

Turgon

The layer 3 vs layer 2 redundancy debate has raged for years. One can design resilience at layer 3 and spanning-tree does have it's headaches. But even the most elegant layer 3 solution can be compromised when someone does something silly like connect a couple of switches together erroneously. Hence I like spanning-tree there as loop avoidance mechanism of last resort

ColbyG

it_consultant wrote: »

Or my personal favorite, "this is spanning tree protocol and this is how we turn it off".

Why would you do that?

it_consultant

Turgon wrote: »

The layer 3 vs layer 2 redundancy debate has raged for years. One can design resilience at layer 3 and spanning-tree does have it's headaches. But even the most elegant layer 3 solution can be compromised when someone does something silly like connect a couple of switches together erroneously. Hence I like spanning-tree there as loop avoidance mechanism of last resort

I have been guilty of plugging in a switchport and seeing the whole stack light up solid. However, the one or two times I have had this problem compared to the many times I have had switchports block for no reason or DHCP sequences which timed out make STP a must for only certain environments where this design is called for.

Turgon

it_consultant wrote: »

I have been guilty of plugging in a switchport and seeing the whole stack light up solid. However, the one or two times I have had this problem compared to the many times I have had switchports block for no reason or DHCP sequences which timed out make STP a must for only certain environments where this design is called for.

One has to account for the environment at hand. If the physical controls are strong, and the number of changes are few, then one goes one way. If there are a myriad of situations that could ensue, some of them out of your control, then you go another. The bottom line is to configure what is *right* to protect you and your customers.

it_consultant

networker050184 wrote: »

Of course there are plenty of simple networks out there, but even with only two routers with some subnets behind them you need a routing protocol. I guess you could go with static routes, but that's not a very good idea. The CCNA teaches you how to set a simple network like this up.

And I won't even get started on turning off STP.

What I've learned from my time working in IT and being on this site is the less you know about a network the more simple you think it is.

I will take issue with the last sentence there. Simply because a network CAN become very complex doesn't mean that it should or that the existence of complexity is an indicator of a well designed network. Most enterprise networks needn't be very complex to run efficiently.

I also find it irritating that the suggestion that since I like to keep things simple, it means that I don't know about the network. It is really the opposite, the more I learn and the more I troubleshoot the more I appreciate keeping things uncomplicated whenever possible.

The addition of a routing protocol does not magically turn a network from simple to complex. In my networks with many remote sites the network is still basically simple. Its not simple for the ISP who is providing me with my VPLS connections; of course that is their specialty.

instant000

it_consultant wrote: »

I will take issue with the last sentence there. Simply because a network CAN become very complex doesn't mean that it should or that the existence of complexity is an indicator of a well designed network. Most enterprise networks needn't be very complex to run efficiently.

it_consultant, I believe that I get your point:

making it more complex doesn't necessarily make it run more efficiently

that's kinda summarized here:

RFC 3439 - Some Internet Architectural Guidelines and Philosophy

At the same time, I get networker050184's point, too, in that networking, even at its most simplest, is still quite complex when you analyze it down to the individual parts.

For an example, look at this RFC on TCP (a protocol you're undoubtedly familiar with using on a daily basis):

http://www.ietf.org/rfc/rfc793.txt

Of course, I could be wrong, and maybe you two should clarify your stances

I'm believing that you're talking about two different issues, is all I'm saying.

Turgon

One should always look for elegance in design. It just makes sense. In terms of spanning-tree, look for IEEE implementations in a mixed environment in terms of vendors, or if you want to be ready for a potentially mixed environment in the future to realise savings for your company. With RSTP, Cisco went their own way with the RPVST+ implementation, so avoid if mixing with non Cisco switches. Look at MST.

it_consultant

instant000 wrote: »

it_consultant, I believe that I get your point:

making it more complex doesn't necessarily make it run more efficiently

that's kinda summarized here:

RFC 3439 - Some Internet Architectural Guidelines and Philosophy

At the same time, I get networker050184's point, too, in that networking, even at its most simplest, is still quite complex when you analyze it down to the individual parts.

For an example, look at this RFC on TCP (a protocol you're undoubtedly familiar with using on a daily basis):

http://www.ietf.org/rfc/rfc793.txt

Of course, I could be wrong, and maybe you two should clarify your stances I'm believing that you're talking about two different issues, is all I'm saying.

My stance, originally, was that CCNA curricula is the best you can get for any kind of network or server professional, but totally lacks the big picture of the networks many of us work on.

gbadman

Am I hallucinating or did Turgon travel through time?;)