Plan 9 from Bell-Labs

Computer Science 400 project by:

Andrey Mirtchovski
Supervised by:
Tony Kusalik

Plan9 from Bell-Labs

Introduction

 An operating system designed by

the people who created UNIX
 Designed with distributed
computing in mind
 Main goals concern performance
and usability in said environment

Plan9 from Bell-Labs

Topics of Discussion
 What is plan9?
 Why plan9?
 Project goals
 Project tentative setup
 Early implementation stages (or
“Having old hardware == trouble”)
 Late implementation stages (or
“Having little time == trouble”)

Plan9 from Bell-Labs

Topics…
 Setup
 Mashines in initial project state
 Machines in adult project state
 Results
 Usability
 Computational fitness
 Conclusions

Plan9 from Bell-Labs

What is Plan9?
 OS
 Created by programmers for
programmers
 Has all the requirements for a full-
blown commercial OS minus
software and price
 Distributed
 Research-Oriented

Plan9 from Bell-Labs

 Full set of development tools
including
 Compiler for several platforms
 Text editors (more on that later)
 Libraries
 GUI
 Full network connectivity
 Open Source

Plan9 from Bell-Labs

Plan9 from Bell-Labs
Typical Plan9 Setup
 CPU
 One or many
 Terminals
 Typically one per user
 Minimal hardware requirements
 File Servers
 Could be incorporated with cpu servers
 Auth Servers
 Security

Plan9 from Bell-Labs

Plan9 from Bell-Labs
Why Plan9?
 Interest in Operating Systems
Design and Implementation
 Creators’ reputation (hence:)
 Quality of code
 Minimalist design
 Code understandability

Plan9 from Bell-Labs

Project goals
 Setup and use a working plan9 research
environment
 Implement and measure fitness of
several Bioinformatics algorithms when
implemented in a distributed manner
 Measure performance and compare to
other (more robust) operating systems
(BSD or Linux)
 Compare ease of use and application
development with other OSs

Plan9 from Bell-Labs

Bioinformatics
 Purpose of bioinformatics
algorithms
 Algorithm used
 Initially decided on BLAST (huge
success in PBI’s research)
 At a later stage moved to FASTA

 Optional other algorithms never

implemented (lack of time)

Plan9 from Bell-Labs

Early stage setup
 The horrors of using old,
unsupported hardware
 Lost tremendous amount of time
within kernel space trying to create
some sort of useful environment so
the project could start
 From of a month (at most) allotted
for initial setup, scope crept to a
whole semester
Plan9 from Bell-Labs
Early stage setup…
 Learning from experience (benefits from
using old hardware):
 Hands got very dirty very early in the
research phase
 Appreciation for minimalist design and
lack of code bloat
 Appreciation for overall OS and hardware
design (serial consoles and cross
compilers)
 Increased understanding of operating
system internals and behaviour

Plan9 from Bell-Labs

Early network organization
 3 CPU servers (udb[123]). Alpha
UDB machines
 Running NetBSD
 Udb2 attempts to boot Plan9
kernel off udb1
 Hard to use…
 Unsuccessful…

Plan9 from Bell-Labs

Late stage setup
 New machines brought thanks to
Supervisor and Dave Bocking
 Luck with hardware (all supported)
 Initial installation and setup of
Plan9 took less than 5 hours, and
installed over the network
 Lack of proper hardware
substituted for lack of time

Plan9 from Bell-Labs

Current network
organization
 3 CPU servers (udb[123])
 CPU servers act as file servers (locally)
 udb1 acts as an authentication server
 All machines can access vital Plan9
services
 All machines allow terminal access from
on-campus machines (using challenge-
response based authentication)
 No root!

Plan9 from Bell-Labs

Results
 Usability
 Extremely programmer-friendly
environment
 Editor created ‘from programmers for
programmers’
 Debugging is a breeze
 Libraries and API exclude code bloat
(unless one is careless)
 IRC client – 200 lines of code
 Graphics – 70 lines of code (bouncing
balls)

Plan9 from Bell-Labs

Results…
 The one true power of shell scripting:
cpu% cat /bin/uptime
#!/bin/rcclock=`{cat /dev/time}
xx=`{{echo $clock(3) / $clock(4);
echo '_/86400'}|hoc|sed 's/^\./0./g;s/\..*//g'}
sec=$xx(1)
days=$xx(2)
cat /dev/sysname
echo -n ' up '$days' days, date $sec |awk
'{print $4}'

Plan9 from Bell-Labs

Plan9 from Bell-Labs
Results…
 Porting
 APE/PSH
 Some programs compile flawlessly

 FASTA

 POVray

 Bad programming practices lead

to unportable/unmaintainable code

Plan9 from Bell-Labs

Results…
 Computational fitness
 Comparable executables speed
with linux running on same
hardware
 Some compiler research has lead
to significant improvement of
execution speed
 Have in mind – operating system is
not tuned for performance (as
Linux/BSD are for example)

Plan9 from Bell-Labs

Results.. POVray
 Accidentally ported to Plan9
 Used to measure raw CPU performance
and compare with Linux
 Speed results:
(create a complex 3d ray-tracing image)
 Linux:
 Total Time: 368 seconds
 Plan9:
 Total Time: 415 seconds

Plan9 from Bell-Labs

Linux Plan9

Image Sizes:
-rw------- 1 aam396 dip 192074 Feb 27 13:46 linux.ppm
-rw------- 1 aam396 dip 192074 Feb 27 13:47 p9.ppm

Plan9 from Bell-Labs

Possible uses of PovRay:
 Rendering farm (tested)
 Rendering single image on several
machines without deploying any
multiprocessor hardware (even though
supported
 Some testing results (180 image
animation):
 povray -- 180 images, single machine:
9311.36u 26.78s 9465.01r
 povray -- 180 images, 3 machines:
3403.79u 10.42s 3251.90r

Plan9 from Bell-Labs

Drawbacks

 Significantly higher network load

then expected
 Very high load imposed on file
server

Plan9 from Bell-Labs

Plan9 from Bell-Labs
Plan9 from Bell-Labs
Bioinformatics
 Algorithm: FASTA, SSEARCH (version
3.3)
 Easily transformable into a distributed
application (break database into smaller
files and search separately – thanks to
support from S. O’Hearn of PBI)
 Several possible distributed scenarios
depending on where the database files are
located – prompted by real problems
encountered at PBI
 Searched the entire ecoli bacteria genome
– bonus points added for fun
Plan9 from Bell-Labs
Possible scenarios:
 Single database on remote
fileserver
 Several small databases on

remote server
(both viable for FASTA and
SSEARCH)
 Locally stored database files

Plan9 from Bell-Labs

Plan9 from Bell-Labs
Some results:
 Algorithm: FASTA (version 3.3)
 Easily transformable into a distributed
application (break database into smaller
files and search separately – thanks to
support from S. O’Hearn of PBI)
 Several possible distributed scenarios
depending on where the database files
are located – prompted by real problems
encountered at PBI
 Searched the entire ecoli bacteria
genome – bonus points added for fun!

Plan9 from Bell-Labs

Results (cont):
 Small database, networked:
 241.68u 0.24s 256.35r run fasta3/fasta33
network
 Small databases, local:
 245.01u 0.31s 248.65r run fasta3/fasta33
local
Difference comes from less context
switching with local database and
corresponds to 3-5 seconds delay in
fetching a 4MB file over a 100baseT
connection.

Plan9 from Bell-Labs

Results (cont):
 Large database, local:
 Scan time: 53.366 Display time:
473.390

Note: local database run time is not

wall-clock execution as above, but
a simple execution of the program.
This does not increase significantly
the result margin.

Plan9 from Bell-Labs

Conclusion
 Bell-Labs research group have presented
us with a one-of-a-kind research platform,
which definitely has the possibility of
attracting students and professors’ attention
 Maybe not a commercially viable product,
but a very good learning tool
 Need to emphasize on the lower learning
curve with regards to deep internals. Code
very easy to read
 Ability to share ideas and learn from some
of the groundbreaking personalities in this
field is an enormous ego boost

Plan9 from Bell-Labs

Would not have been be
possible without:

Tony Kusalik (CS)

Stephen O’Hearn (PBI)
Dave Bocking (CS)
Cary Bernath and Greg Oster (CS)

Plan9 from Bell-Labs

Questions?

Plan9 from Bell-Labs