/bb|[^b]{2}/
Never stop Grokking
Wednesday, December 12, 2012
A correlation between load time and usage
We frequently see reports of website usage going up as load time goes down, or vice-versa. It seems logical. Users use a site much more if it's fast, and less if it's slow.
However, consider the converse too. Is it possible that a site merely appears to be faster because users are using it more, and therefore have more of it cached? I've seen sites where the server-side cache-hit ratio is much higher when usage is high resulting in lower latency.
At this point I haven't seen any data that can convince me one way or the other. Do you?
Labels
2fa
4.01-strict
404
accessibility
acer
airport wifi
algorithm
android
apache
API
att
audio
australia
authentication
ayttm
badges
bandwidth
bbc
bcp
berlin
bigsur
blog
blogger
blogger template
bof
book
boomerang
broken
bug
byte order
c
caching
chrome
closure
cmc
cms
codepo8
colours
comic strip
comments
communication
compile
conference
confoo
congestion-control
cookies
correlation
couchdb
cracker
crash
creative
credit card
crockford
cron
crux
csrf
css
curl
cwv
data tags
database
date
db
delicious
design
developer
dhtml
dns
docker
dom
dopplr
dragdrop
DST
dynamic script node
education
email
emotion
endianness
entities
ephemeral ports
epicondylitis
error checking
esmtp
everybuddy
extensions
facebook
favicon
fc9
fedora
fidelity
filesystem
firefox
firesheep
flickr
flot
form
forms
fosdem
foss
foss.in
freebsd
freedom
freestyle
ftp
function currying
gdb
geek
geo
george
gmail
gnome
google
gradient
groupon
hack
hacker
hardy
hash
howtos
htc
html
html5
http
i18n
icici
ie
iit
im
innerHTML
instant messaging
interfaces
internet
internet explorer
ios
ip
ip address
ipc
iphone
ipv6
iso8601
ISP
jabber
javascript
jslint
json
julia
keynote
latency
latex
LC_TIME
lcp
linux
load
localisation
login
lsm
luhn
MAC
macosx
mail
mathematics
mathjax
measurement
media queries
meetup
memory
messaging
microformats
missing kids
mobile
montreal
movable type
mp3
mvc
mysql
name generator
navtiming
network
nexus
nodejs
notes
opensource
opera
partition
passwords
pdf
perception
performance
perl
phone
php
planet
png
ports
prerender
printing
privacy
programming
programming style
progressive enhancement
psychology
puzzle
recovery
redhat
regex
regular expressions
resource timing
review
rfc2822
rfc3339
rhel
roundtrip
rss
rum
rwd
safari
sampling
scalability
scripting
search
secnet
security
sed
segfault
self extracting tarball
sendmail
seo
server
shell
shell script
sigdashes
site
slideshare
smtp
soap
sockets
spoofing
SQL
ssl
starttls
startup
statistics
stoyan
strftime
stubbornella
sydney
sysadmin
tablespace
talks
tcp
tcp-slow-start
testing
theme
thisisbroken
thisisfixed
thoughts
throughput
tim berners-lee
timezone
tips
toc
toy
transactions
twitter
two factor auth
typing
ubuntu
ui
unicode
unix
url
usability
ux
velocity
vint cerf
w3c
wav
web
web services
webcam
webdev
webdu
webkit
webperf
webtiming
whois
widgets
wifi
windows
workaround
write performance
X
xss
yahoo
ydn
YQL
yslow
yui
Translate this page
- PHOTO FEED
- Blog feed bluesmoon@Mastodon
- © PHILIP TELLIS
The other side of the moon by Philip Tellis is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License.
6 comments :
It seems to me like there could be two different types of sites that would have different explanations for this observation. I can imagine one type of site where you really would need lots of users hitting the site in order to cache much of the site. This site would have a lot of different content depending on user interaction or content that changes frequently. In those conditions, high traffic would be necessary to cache much of the site. Then, you might draw the conclusion that the high traffic is reducing load times because more of the site is cached.
But what about a different site that doesn't require a lot of users to cache most of its content. Maybe its content is relatively static or maybe most of its content, by nature, can't be cached (e.g. financial data during market hours). If you were seeing high traffic and low load times, then I think your first idea could apply: users are using the site because its fast.
What do you think? Does that make sense?
Hi Jared, thanks for your comment.
Yes, I do think both types of sites exist, and there's a whole spectrum in between. For example, in the financial data site you mention, there could still be many static assets (JavaScript, CSS, Images, etc.) that are cached by the user's browser. In that case even a few users hitting the site many times will make it appear fast.
What we need is a set of metrics that help us determine what kind of site we have an which of the two is the cause and which is the effect.
It's hard to determine human behaviour when you can neither see nor speak to the humans in question.
Yes that is definitely possible. More users using the site from various locations will warm up not only their own browser cache but also caches on various CDN instances replicating the site's static resources. Thus, the experience will be better. The site will appear faster even without any specific code changes.
Another point to make is that when a site is loading quickly, you have the ABILITY to make more requests because you aren't waiting around as much.
what about high usage lead to high latency(high load time) ? like irctc.co.in but still people use it to book the tickets.
@namma yes, we see many sites like that. Performance is irrelevant for them because it's the only place to get a particular service. In a sense, it's a monopoly and can do/charge anything it wants without fear of losing customers.
Post a Comment