Re: Programmers with network engineering skills

[Alain Hebert <ahebert () pubnix net>]

    About (5 thru 6)

    Hard to keep a straight face in front of a customer when, after 
assigning him a IP in our 192.172.250.0 range...

    ... He ask why are we NATing using private IP's.

    We also had plenty of experience with ppl getting confused about 
16, 17.

    Your could add L2 Trunking and VRRP to your list...  I spent many 
hours explaining those to no avail on many occasion.

    Sad.

-----
Alain Hebert                                ahebert () pubnix net
PubNIX Inc.
50 boul. St-Charles
P.O. Box 26770     Beaconsfield, Quebec     H9W 6G7
Tel: 514-990-5911  http://www.pubnix.net    Fax: 514-990-9443


On 03/05/12 21:36, Jimmy Hess wrote:
> On Mon, Mar 5, 2012 at 6:09 PM, Justin M. Streiner
> <streiner () cluebyfour org>  wrote:
>
> > Admittedly we (the 'network guys') don't always make it easy for them. RFCs
> > get obsoleted by newer RFCs, but the newer RFCs might still reference items
> > from the original RFC, etc.  This can turn into developing for something
> Yes, this is problematic.    The preferred result should be one specification
> for each protocol,   with references only for optional extensions.
>
> > Other common, but misguided assumptions (even in 2012):
> > 1. You will be using IPv4.  We have no idea what this IPv6 nonsense is.
> > Looks complicated and scary.
> > 2. 255.255.255.0 is the only valid netmask.
> > 3. You are using Internet Explorer, and our web management interface has
> > ActiveX controls that require you to do so.
> > 4. You will be assimilated.  Resistance is futile.
> Add some additional misguided assumptions:
>
>     (5)  Any IP address whose first octet is 192.  or  1.  is a private IP.
>     (6)  Any IP address whose first octet is not 192.  is not a valid LAN IP.
>     (7)  Any IP address whose last octet is .0  is an invalid IP host address
>     (8)  Any IP address whose last octet is .255 is an invalid IP host address
>
>     (9)  If my DNS service supports DNSSEC validation, even with no trust anchors
>           configured,  it's cool to go ahead and send all queries with
> the CD and DO bits
>           set to 1
>           and perform no validation;  it's even cooler if I only
> support SHA1 keys and
>           no RSA/SHA-256.
>
>    (10)  Everyone enters their NTP,  and AD servers by IP address, so it
>           is best to  have a textbox that only allows IPs,  not hostnames.
>
>    (11)  Nobody actually uses SRV records, so don't bother looking for them.
>
>    (12)  Once a DNS lookup has been performed, the IP never changes, so
> it makes sense
>           to keep this in memory  until we reboot.
>
>    (13)  Nobody has more than 1 recursive DNS server,  1 NTP server, 1
> LDAP server,
>           1 Syslog server,  and  1 Snmp management station;
>           so a single IP entry text box  for each will suffice.
>
>    (14)  Nobody has more than 2 recursive DNS servers, so just allow
> only 2 to be entered.
>
>    (15) 30 seconds per resolver seems like a good timeout for DNS queries, so no
>          need for a configurable timeout;  just  try each server
> sequentially, make the
>          UI hang, the user will be happy to wait 5 minutes;  also make
> the service
>          provided by the device temporarily stop --   users likes it
> when their devices
>          stop working, to remind them to get their first DNS server back up.
>
>     (16)  The default gateway's IP address is always 192.168.0.1
>     (17) The user portion of E-mail addresses never contain special
> characters like  "-" "+"  "$"   "~"  "."  ",", "[",  "]"
>
>
>
> > jms
> --
> -JH