Great Circle Associates Firewalls
(April 1996)
 

Indexed By Date: [Previous] [Next] Indexed By Thread: [Previous] [Next]

Subject: Re: Most popular firewall implementation?
From: me @ tartufo . muc . ditec . de (Michael Elbel)
Date: Mon, 29 Apr 96 10:41 MSZ
To: genel @ inforamp . net
Cc: firewalls @ greatcircle . com
Newsgroups: comp.security.firewalls
References: <4lm7pv$ifs @ hudson . cs . columbia . edu> <4luqlh$t7m @ news . inforamp . net> <31831C99 . 75CB @ devine . net> <4m08a6$b2h @ news . inforamp . net>
Reply-to: me @ mail . muc . ditec . de

In comp.security.firewalls you write:

>Not disputing the fact, but I would be concerned about scalability issues. 
>From what I understand, Borderware is a PC solution. A PentiumPro may only be 
>able to handle that much load as an enterprise-wide gateway, whereas a 
>firewall run on a scalable platform (ie. RS/6000, SunSparc, DEC) will probably 
>offer more horsepower and upgradeability.

I don't know too much about how the software scales (OS, Firewall SW
itself) for Borderware, but it looks like reasonably modern PC-based
hardware can compete pretty well in the internet server business,
shoving data around at speeds faster than Ethernet. Take Walnut Creek
CDROM's ftp and www server:

Quoting config from WC's documentation 
(the whole thing is on ftp://cdrom.com/config):

----
wcarchive.cdrom.com is an Intel architecture PC machine running the FreeBSD
operating system.

[...]

One 150Mhz P6 CPU ("Pentium Pro")
512MB of main memory (72pin, 16M x 36bits x 60nsec SIMMS X 8)
3 Adaptec AHA-2940 PCI SCSI controllers
1 SMC 9332 (DEC DC21140 based) PCI 100Mbps Fast Ethernet controller
18 attached SCSI drives of various types, typically 4.3GB.
   These are all standard SCSI-II drives, all from Quantum and Seagate.
   No special RAID controller hardware or WIDE drives are used.

[...]

Our connection to the Internet is via a 100Mbps Fast Ethernet channel that
connects directly to a core router which in turn connects to several major
points on the Internet via multiple T3s.

More than 25,000 people visit wcarchive each day, and wcarchive sends out
more than 2 terabytes of files each month (as of April, 1996), limited mostly
by the Internet backbone.

[...]
------

Now some data Jordan Hubbard posted this spring:

I thought people would be amused to see the load at the >1000 user
scenario:

jkh @
 wcarchive-> ftpcount
Service class mirror-ftpserv       -   0 users ( 10 maximum)
Service class mirror-freebsd       -   0 users ( 10 maximum)
Service class mirror-linux         -   0 users ( 10 maximum)
Service class mirror-demos         -   0 users ( 10 maximum)
Service class mirror-os2           -   0 users ( 10 maximum)
Service class mirror-upl           -   0 users ( 10 maximum)
Service class mirror-test          -   0 users ( 10 maximum)
Service class local                -   0 users ( 10 maximum)
Service class remote               -   2 users ( 10 maximum)
Service class anonymous            - 1012 users (1250 maximum)

jkh @
 wcarchive-> top
load averages: 12.18,  9.51,  7.99                                    14:48:20
1128 processes:9 running, 1118 sleeping, 1 zombie
Cpu states: 13.5% user,  0.0% nice, 27.3% system, 19.9% interrupt, 39.4% idle
Memory: 307M Active, 2356K Inact, 50M Wired, 129M Cache, 676K Free
Swap:   819M Total, 804M Free, 2% Inuse  

  PID USERNAME PRI NICE   SIZE   RES STATE   TIME   WCPU    CPU COMMAND
   82 root       2    0   180K  268K sleep 300:12  2.71%  2.71% syslogd
23487 root       2    0   464K  320K sleep   0:15  1.83%  1.83% ls
22605 dave       2    0  7324K 7636K sleep   0:31  1.18%  1.18% perl
28761 jkh       85    0  1464K 1596K run     0:00  2.19%  0.65% top
28176 root       2    0   700K  460K sleep   0:00  0.54%  0.53% ftpd
28775 root       2    0   592K  448K sleep   0:00  3.83%  0.53% ls
28772 root       2    0   700K  440K sleep   0:00  2.31%  0.42% ftpd
28773 root       2    0   680K  428K sleep   0:00  2.31%  0.42% ftpd
28784 root       2    0   680K  424K sleep   0:00  8.59%  0.42% ftpd
28754 root       2    0   680K  440K sleep   0:00  1.04%  0.34% ftpd
28714 root       2    0   700K  448K sleep   0:00  0.61%  0.31% ftpd
28749 root       2    0   680K  440K sleep   0:00  0.92%  0.31% ftpd
28710 root       2    0   680K  424K sleep   0:00  0.58%  0.31% ftpd
26718 www        2    0   684K  440K sleep   0:00  0.27%  0.27% httpd
28763 root       2    0   680K  428K sleep   0:00  1.03%  0.27% ftpd

Sometimes this machine just amazes me.. :-)

						Jordan
----

I know this kind of serving can't be directly compared to the type of
load a firewall gateway takes, there's issues of actual hardware
reliability, maintainability, *ility, etc. But dismissing any solution
solely upon one aspect of a system such as hardware is IMO not the way
to go.


>Note: I really hope this doesn't become a "My-Firewall-Is-Better-Than-Your-
>Firewall" discussion. Different solutions for different requirements...

Nah, merely a "my-dick-is-larger-than-yours" one.

Michael
-- 
Michael Elbel, DITEC, Muenchen, Germany - me @
 muc .
 ditec .
 de
Fermentation fault (coors dumped)

Indexed By Date Previous: Dual Homed Host and Proxy services
From: kwakh @ ctrvax . Vanderbilt . Edu (Joon)
Next: Firewall for LINUX
From: ehoiting @ se-rijswijk . nl (Erik_Hoitinga)
Indexed By Thread Previous: Re: Dual Homed Host and Proxy services
From: Duan Zhenhai <ccdzh @ shoukui . pku . edu . cn>
Next: Re: Most popular firewall implementation?
From: todd @ momentum . com . au (Todd Hooper)

Google
 
Search Internet Search www.greatcircle.com