2. A Self Introduction
Been involved with technology since 1981
Led development of GoldMine for 4 years
CTO of Famundo, LLC
Contributed libraries and educational
materials as open source
Currently CTO for Morph Labs, Inc.
3. The Need to Scale
Future is Uncertain
Save Money and Effort
Optimize Risk/Expenditure
Time is Short
4. What can we scale?
Web Servers
Application Servers
Static Files
Storage
Databases
5. What technologies should
you use?
Simple to start with
Able to grow
Easy to replace
Know what you are doing
8. How do you plan scaling?
No “one size-fits-all” solution -
depends on the specific application
Check, measure, test and repeat
Start from real bottlenecks - not from
assumptions
Choose the path of least resistance!
9. Easy vs. Complex
Do you know what’s easy?
Better code and caching covers 90% of
scaling needs!
Web servers
Application servers
Messaging
Storage
Database
10. Don’t overdo it
Do you really need the “it”?
?
20 machine DB cluster?
CDN for 10,000 pages
a day?
Mirrored Messaging
servers?
11. Future Proofing?
Can you estimate your future needs now?
Do you have time to test and analyze all
scenarios?
Proven technology
Embrace Change
Repeatability - Testing
12. Case Study:
Tracks @ Morph
Open source app that implements
GTD
Originally stand-alone, now its a
multi-tenant SaaS app (c/o Morph)
iPhone Mobile Safari compatible
Thousand of users since we
launched it (mid 2008)