A Silicon Valley-like startup... in Spain :)
J. Javier Maestro – [email protected]
Backend Engineer – Core & Scalability Team
Slide 2
Slide 2 text
Tuenti is...
Slide 3
Slide 3 text
No content
Slide 4
Slide 4 text
No content
Slide 5
Slide 5 text
No content
Slide 6
Slide 6 text
No content
Slide 7
Slide 7 text
No content
Slide 8
Slide 8 text
No content
Slide 9
Slide 9 text
Now, seriously...
* the TV, Rockband and Paintball team building are real, though! :)
Slide 10
Slide 10 text
2006
Slide 11
Slide 11 text
No content
Slide 12
Slide 12 text
No content
Slide 13
Slide 13 text
Today
Slide 14
Slide 14 text
No content
Slide 15
Slide 15 text
No content
Slide 16
Slide 16 text
No content
Slide 17
Slide 17 text
No content
Slide 18
Slide 18 text
+200 employees
65% engineers
18 nationalities
Slide 19
Slide 19 text
Back to the
Future?
Slide 20
Slide 20 text
No content
Slide 21
Slide 21 text
Tuenti Facts
Slide 22
Slide 22 text
BIG Numbers
+10 million users
+150 million daily chat messages
+4 million daily uploaded photos
Slide 23
Slide 23 text
BIG Numbers
+30 billion monthly page views
+27k web requests/second
+12Gbps peaks (including CDNs)
+4Gbps just the datacenter
+1000 servers
Slide 24
Slide 24 text
BIG Numbers
50% of users on site daily, averaging
3 sessions per day
82 minutes per day
20% of users use mobile features
daily
Slide 25
Slide 25 text
La palabra más buscada en Google España
(Google Zeitgeist 2009 y 2010 YTD)
Slide 26
Slide 26 text
No content
Slide 27
Slide 27 text
No content
Slide 28
Slide 28 text
No content
Slide 29
Slide 29 text
“The Spanish friendship map” by Barcelona Media
Slide 30
Slide 30 text
No content
Slide 31
Slide 31 text
Viral
Slide 32
Slide 32 text
No content
Slide 33
Slide 33 text
No content
Slide 34
Slide 34 text
No content
Slide 35
Slide 35 text
No content
Slide 36
Slide 36 text
No content
Slide 37
Slide 37 text
Tech Teams
Slide 38
Slide 38 text
Frontend
Slide 39
Slide 39 text
No content
Slide 40
Slide 40 text
Mobile
Slide 41
Slide 41 text
No content
Slide 42
Slide 42 text
Architecture
Slide 43
Slide 43 text
No content
Slide 44
Slide 44 text
Dev Tools
Slide 45
Slide 45 text
No content
Slide 46
Slide 46 text
Testing
Framework
Slide 47
Slide 47 text
No content
Slide 48
Slide 48 text
QA
Quality Assurance
Slide 49
Slide 49 text
No content
Slide 50
Slide 50 text
Systems
Slide 51
Slide 51 text
No content
Slide 52
Slide 52 text
IT (crowd :)
Slide 53
Slide 53 text
No content
Slide 54
Slide 54 text
Design & UX
Slide 55
Slide 55 text
No content
Slide 56
Slide 56 text
Backend
Slide 57
Slide 57 text
No content
Slide 58
Slide 58 text
+much more!
User Support
Sales, Marketing
Law, Finance
Office Staff
Cleaning Staff
Slide 59
Slide 59 text
Magic stuff...
Slide 60
Slide 60 text
In the beginning…
One or few servers for the
entire application
Monolithic structure
Bottleneck: Database reads
Slide 61
Slide 61 text
Cache
Switch out heavy DB reads
for light cache reads
Significant reduction in DB
load
Typical page load has
under 2 or 3 queries, and
up to hundreds of cache
reads
Bottleneck remains the
same as before
Slide 62
Slide 62 text
Replication
Master / Slave
architecture
Split up reads and
writes
Optimized
configurations
Bottleneck: dataset
size
Slide 63
Slide 63 text
Data Separation
Divide your schema into
logical blocks
One cluster per data type
More complicated to
manage in code.
Bottleneck: Overload on a
given data type
Slide 64
Slide 64 text
Data Partitioning
Divide a single table into
parts and spread them
over several clusters
Almost limitless
scalability
Not free – Code
complexity increases
significantly
Bottleneck: dataset size
(again)
Slide 65
Slide 65 text
Cache Partitioning
Partitioning does not
need to be limited to
databases
Partitioned
Memcached and
frontend farms
Improves internal
network usage
Cross partitioning
scheme between
cache and database
Slide 66
Slide 66 text
Archiving
Storing less-frequently-
used data in alternative
storage
Keeps main table size
within known limits
Flexibility when
choosing alternative
storage
Slide 67
Slide 67 text
Client-side routing
Use browser logic to load balance over
frontend farms
Reduce server load
Remove a potential SPOF
Slide 68
Slide 68 text
Content Delivery Networks
Static content
Optimized for massive scaling
Everything has a price – choose your
optimizations wisely