Re: [Misc][Rant] Internet router (straying slightly OT)
On 9/29/05, Warren Kumari <warren@kumari.net> wrote:
I have met "Senior Network Engineers" who don't understand longest match rule ("The traffic will take 10/8 instead of 10.0.0.0/24 because it has a better admin distance", "I can override these 300 OSPF routes with a single static supernet", etc), who believe that routers will not route between directly connected interfaces without putting them into a routing protocol, that transit networks don't need a full mesh of iBGP[1] because "you can just redistribute BGP into [OSPF/IS-IS/IGP of choice], that ICMP uses TCP as a transport, etc.
In a similar note, I Do care about networks and the like but fail to fully understand the extensive details of how it all works. I do not proclaim myself to be an engineer and try to stick with what I do well. I read rfc, wikipedia, etc but just don't know what /to/ read. I had never heard of iBGP, OSPF, IS-IS untill today. What I need, and I'm sure quite a few others who listen to this list for insight, is a good reference to pick up and read that will cover said topics and beyond. I finally got the basic concept to CIDRs and how they work thanks to this list and Google. I know this message is slightly off topic from NANOG, but kinda fits in response to parent and am hoping not to get flamed. Any suggestions? A Padawan, Mark Owen
I'd start with Sam Halabi's Internet Routing Architectures book.
In a similar note, I Do care about networks and the like but fail to fully understand the extensive details of how it all works. I do not proclaim myself to be an engineer and try to stick with what I do well. I read rfc, wikipedia, etc but just don't know what /to/ read. I had never heard of iBGP, OSPF, IS-IS untill today. What I need, and I'm sure quite a few others who listen to this list for insight, is a good reference to pick up and read that will cover said topics and beyond. I finally got the basic concept to CIDRs and how they work thanks to this list and Google.
I know this message is slightly off topic from NANOG, but kinda fits in response to parent and am hoping not to get flamed.
Any suggestions?
A Padawan, Mark Owen
On Sep 29, 10:42pm neil@DOMINO.ORG wrote:
I'd start with Sam Halabi's Internet Routing Architectures book.
Grumble ... with reference to the issue of routing between connected networks, your choice might be too advanced. Douglas Comer's "Internetworking With TCP/IP Vol I" has been one of the best introductions for years (soon literally "decades"), and should be required reading before progressing to more advanced topics. Life begins with ARP. http://www.cs.purdue.edu/homes/dec/netbooks.html Best, -- Per
On 9/29/05, Mark Owen <mr.markowen@gmail.com> wrote:
Any suggestions?
Keep reading everything you can get your hands on. When faced with a question like "who owns this router?", don't waste your time signing up for a mailing list just to make a fool of yourself. Do some research. Keep reading. And before you know it, you'll have taught yourself an amazing amount of knowledge. It really is that simple.
Any suggestions?
Keep reading everything you can get your hands on. When faced with a question like "who owns this router?", don't waste your time signing up for a mailing list just to make a fool of yourself. Do some research. Keep reading. And before you know it, you'll have taught yourself an amazing amount of knowledge.
It really is that simple.
Alternatively, force yourself to study for Cisco's CCNA. That will, at the very least, give you a basic (vendor-specific?) understanding of networking. -- matthew zeier - "Curiosity is a willing, a proud, an eager confession of ignorance." - Leonard Rubenstein
On Thu, Sep 29, 2005 at 02:45:20PM -0700, Aaron is rumored to have said:
When faced with a question like "who owns this router?", don't waste your time signing up for a mailing list just to make a fool of yourself. Do some research. Keep reading. And before you know it, you'll have taught yourself an amazing amount of knowledge.
I'm not a guru like most on this list, which is why I rarely chime in despite having been subscribed for 5 or 6 years, but Aaron nailed it here. I used this technique when I wanted to run my own DNS, mail & web servers and managed to teach myself all I needed to know to accomplish those things simply by reading what was available online. I also picked up a few books along the way, but everything you need is on the 'net. St- -- Tower, Observatory You stand in what appears to be a room designed specifically for studying the stars. Lying haphazardly cast about the room are several charts with constellations, some neatly folded, others lying left open. Beside a few of the starcharts are several magical tomes.
On Thu, Sep 29, 2005 at 05:39:30PM -0400, Mark Owen wrote:
Any suggestions?
Start with the OSI[1] model to grasp the fundamentals, next make sure you have a basic knowledge of how TCP/IP addressing works[2]. To get an understanding of routing-protocols, begin with RIP[3] and perhaps run your own RIP-lab by using Quagga[4] software on a Linux box. That will get you off the streets for a weekend :) Cheers, -- Sabri please do not throw salami pizza away [1] http://www.cisco.com/univercd/cc/td/doc/cisintwk/ito_doc/introint.htm [2] http://www.networkclue.com/routing/tcpip/addressing.php [3] http://www.livinginternet.com/i/iw_route_igp_rip.htm [4] http://www.quagga.net
On Fri, 30 Sep 2005 05:50:52 EDT, "Robert E.Seastrom" said:
Sabri Berisha <sabri@cluecentral.net> writes:
To get an understanding of routing-protocols, begin with RIP[3] and perhaps run your own RIP-lab
necromancy will be severely punished.
Sayeth RFC1925: (4) Some things in life can never be fully appreciated nor understood unless experienced firsthand. Some things in networking can never be fully understood by someone who neither builds commercial networking equipment nor runs an operational network. Just remember, all you dabblers - a properly designed lab environment is called for, for the same reasons a pentagram is called for... :)
To get an understanding of routing-protocols, begin with RIP[3] and perhaps run your own RIP-lab necromancy will be severely punished.
many hand-on routing workshops start with rip, though with the warning "you will now learn why not to use rip." it makes it easy to teach poison reverse, ... in a relatively small setting. randy
On 30-Sep-2005, at 09:32, Randy Bush wrote:
To get an understanding of routing-protocols, begin with RIP[3] and perhaps run your own RIP-lab
necromancy will be severely punished.
many hand-on routing workshops start with rip, though with the warning "you will now learn why not to use rip." it makes it easy to teach poison reverse, ... in a relatively small setting.
RIP also has the advantage that a worked, non-trivial example of the protocol can fit on a whiteboard, which makes it a reasonable way to teach the concept of a routing protocol to a classroom full of people who have never heard of such at thing. Absolutely agreed, however, that such teaching also necessarily involves emphatic shouting of "YOU WILL NOT TURN THIS ON IN YOUR PRODUCTION NETWORK". [ObAnecdote: I once heard of an airline reservations desk in Hong Kong which had a backup connection to the airline's main centre of operations far distant from Hong Kong, using dial-on-demand ISDN, circa 1995. The monthly invoice for international ISDN charges that followed a contractor's decision to "fix the router by turning on RIP" was apparently an impressive thing to behold, especially given the agressive ISDN idle tear-down configured on the router and minimum 1-minute billing per call.] Joe
On Fri, Sep 30, 2005 at 10:01:34AM -0400, Joe Abley wrote: Hi,
RIP also has the advantage that a worked, non-trivial example of the protocol can fit on a whiteboard, which makes it a reasonable way to teach the concept of a routing protocol to a classroom full of people who have never heard of such at thing.
Which is exactly the reason why I mentioned RIP as a routing protocol to start with. Using RIP instead of OSPF or IS-IS has 2 advantages: one is the simplyness of the concept and the second one you already mentioned:
Absolutely agreed, however, that such teaching also necessarily involves emphatic shouting of "YOU WILL NOT TURN THIS ON IN YOUR PRODUCTION NETWORK".
You learn why not to use RIP in an early stage of your career. Mentioning the terms "router-lsa", "network-summary-lsa" or "nssa-lsa" to a person who potentially does not even know the difference between a distance-vector and a link-state protocol has no positive effect on the learning curve. -- Sabri please do not throw salami pizza away
In message <17213.15974.330499.500268@roam.psg.com>, Randy Bush writes:
To get an understanding of routing-protocols, begin with RIP[3] and perhaps run your own RIP-lab necromancy will be severely punished.
many hand-on routing workshops start with rip, though with the warning "you will now learn why not to use rip." it makes it easy to teach poison reverse, ... in a relatively small setting.
And it's much easier to understand, at least for a beginner. --Steven M. Bellovin, http://www.cs.columbia.edu/~smb
On Sep 30, 2005, at 11:53 AM, Steven M. Bellovin wrote:
In message <17213.15974.330499.500268@roam.psg.com>, Randy Bush writes:
To get an understanding of routing-protocols, begin with RIP[3] and perhaps run your own RIP-lab
necromancy will be severely punished.
many hand-on routing workshops start with rip, though with the warning "you will now learn why not to use rip." it makes it easy to teach poison reverse, ... in a relatively small setting.
And it's much easier to understand, at least for a beginner.
I've been teaching routing protocols for a long time, and I almost always start with RIP for people who don't know what "protocol" means. If you start with OSPF or IS-IS, you invariably get caught up in things like "what is 'link state'?" or "why does Shortest Path First take that path when it's 'longer'?" Plus any "Internet" engineer needs to know about things like hop-count before they can truly understand BGP. Also, I usually include reasons to use RIP in a production network. They are few and far between, but RIP has properties which could be considered "features" when compared to other protocols. -- TTFN, patrick
On Thu, 29 Sep 2005 17:39:30 -0400 Mark Owen <mr.markowen@gmail.com> wrote:
On 9/29/05, Warren Kumari <warren@kumari.net> wrote:
I have met "Senior Network Engineers" who don't understand longest match rule ("The traffic will take 10/8 instead of 10.0.0.0/24 because it has a better admin distance", "I can override these 300 OSPF routes with a single static supernet", etc), who believe that routers will not route between directly connected interfaces without putting them into a routing protocol, that transit networks don't need a full mesh of iBGP[1] because "you can just redistribute BGP into [OSPF/IS-IS/IGP of choice], that ICMP uses TCP as a transport, etc.
In a similar note, I Do care about networks and the like but fail to fully understand the extensive details of how it all works. I do not proclaim myself to be an engineer and try to stick with what I do well. I read rfc, wikipedia, etc but just don't know what /to/ read. I had never heard of iBGP, OSPF, IS-IS untill today. What I need, and I'm sure quite a few others who listen to this list for insight, is a good reference to pick up and read that will cover said topics and beyond. I finally got the basic concept to CIDRs and how they work thanks to this list and Google.
I know this message is slightly off topic from NANOG, but kinda fits in response to parent and am hoping not to get flamed.
FWIW, I would suggest ISP Survival Guide: Strategies for Running a Competitive ISP (Paperback) by Geoff Huston - ISBN: 0471314994, which does a good job with the basics, and is pretty easy to read. Regards Marshall Eubanks
Any suggestions?
A Padawan, Mark Owen
participants (15)
-
Aaron Glenn -
Joe Abley -
Mark Owen -
Marshall Eubanks -
matthew zeier -
Neil J. McRae -
Patrick W. Gilmore -
Per Gregers Bilse -
Petri Helenius -
Randy Bush -
Robert E.Seastrom -
Sabri Berisha -
Steve Thomas -
Steven M. Bellovin -
Valdis.Kletnieks＠vt.edu