Page:1 © 2006 Bala Iyer (bala@bu.edu) ,
Page:1 © 2006 Bala Iyer (bala@bu.edu),. Page:2 © 2006 Bala Iyer (bala@bu.edu) Questions What is Google? What are Google’s core-capabilities? How would.
Dec 25, 2015
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page:1 © 2006 Bala Iyer (bala@bu.edu)
,
Page:2 © 2006 Bala Iyer (bala@bu.edu)
Questions
What is Google?What are Google’s core-capabilities?How would you use some of Google’s capabilities?Business principles
Page:3 © 2006 Bala Iyer (bala@bu.edu)
Competitor response
Law suit – Dr. Kai-Fu LeeTalking to AOLReorg and Vista redesignAdCenter in 2005Oracle busy with acquisition digestionSun forms partnershipYahoo!
Page:4 © 2006 Bala Iyer (bala@bu.edu)
Web 2.0Source: http://www.oreillynet.com/pub/a/oreilly/tim/news/2005/09/30/what-is-web-20.html
• Program to program Program to people• Components everywhere• designed vs. emergence• do-it-yourself
Page:5 © 2006 Bala Iyer (bala@bu.edu)
Industry Stack
Systems(OS, file sharing, terminal access, db access, client-server)
Middleware Services(systems management, application & transaction servers, collaboration & messaging
database)
Application software(CRM package, SCM package, BI, HR, engineering & design, e-biz)
Services(IT consulting, systems integration, outsourcing, training & education,
maintenance)
Data management(meta data, distribution, schema evolution, temporal)
Hardware
Page:6 © 2006 Bala Iyer (bala@bu.edu)
Hardware
Own hardware configuration/cluster ~100, 000 servers
> 350 racks >30,000 m/c >60,000 CPUs 126,368 Ghz of processing power 63,184 GB of RAM 2,527 TB of hard disk space from Prof.
Strassman’s lecture Scale servers while reduce cost/CPU cycle
Page:7 © 2006 Bala Iyer (bala@bu.edu)
Query processing architecture
Each query reads 100s of Mb of data, 1000+ servers and 10s of billionsof CPU cycles. 150 million queries/day (2000/sec); response time < 0.25 sec
Each cluster has the same architecture
8 billion web pages, 1B+ images, now messagesdistributed over clusters
Google web server
Index serverIndex serverIndex serverIndex servers
Index serverIndex serverIndex serverDocument servers
Information about users Ad servers
TB of data
Page:8 © 2006 Bala Iyer (bala@bu.edu)
Operating systems
Custom developed Distributed OS Google file system (GFS)
Metadata management Data dictionary and markup language for
interoperability Global work queue (GWQ) MapReduce
Real-time coordination; workload distribution; fault tolerance; performance monitoring
Based on LinuxJava (recent announcement)Desktop bar
Page:9 © 2006 Bala Iyer (bala@bu.edu)
Data
UserDesktop filesCookieKey wordsGmail 8 billion pages (recently claimed to have three times that of the closest competitor)
Page:10 © 2006 Bala Iyer (bala@bu.edu)
Tools & applications
Core apps AdWords, AdSense Google Earth Search Froogle Google Talk StarOffice (52
million downloads)
Page:11 © 2006 Bala Iyer (bala@bu.edu)
Page:12 © 2006 Bala Iyer (bala@bu.edu)
Services
GmailGoogle MiniApplication provisioning “dark fiber” “Wi-Fi”
APIs
Feb 2006Adaptors for:DocumentumFileNet, Oracleand Siebel
PageRank + GSA
Web data
Enterprise data
Page:13 © 2006 Bala Iyer (bala@bu.edu)
APIs everywhere!
For Rent Powered by craigslist and Google Maps
Page:14 © 2006 Bala Iyer (bala@bu.edu)
Mash-ups everywhere!
0.52 betweeness centrality0.78 structural hole
Page:15 © 2006 Bala Iyer (bala@bu.edu)
Page:16 © 2006 Bala Iyer (bala@bu.edu)
Page:17 © 2006 Bala Iyer (bala@bu.edu)
Page:18 © 2006 Bala Iyer (bala@bu.edu)
How does it work?
AuctionsSatellite
data
AJAX
API
Every call is tacked
Ebay Amazon Yahoo!
Pricing
Case by case (1000/day free)
Individual 10,000/mo freeBasic 30,000/mo free $2.90/1000 queries + $500 Professional $1.6/1000 queries + $1000Enterprise $1.25/1000 queries + $5000
Free (10,000/day)
Free
data
Page:19 © 2006 Bala Iyer (bala@bu.edu)
Business models
Advertising Affiliate programs/lead generationSubscriptionsPay-per-transaction
Page:20 © 2006 Bala Iyer (bala@bu.edu)
The old order changeth, giving place to new
Source: ZDNET http://blogs.zdnet.com/BTL/?p=2484Mashup ecosystem poised to explode Posted by David Berlind
Page:21 © 2006 Bala Iyer (bala@bu.edu)
Old stack metaphor
Windows, Linux, MacOSOS
AP
I Lay
er
Net
wor
king
Gra
phic
s
Sto
rag
e
Sec
urity
Mes
sagi
ng
Mem
ory
App
s La
yer
Gam
es
Map
ping
Cal
enda
ring
Tra
nsac
tion
proc
Off
ice
pro
duct
ivity
Soc
ial n
etw
orki
ng
Physical assets
Env
iron
men
t
Managing computer resources
Page:22 © 2006 Bala Iyer (bala@bu.edu)
New services stack
Internet/Web 2.0OS
AP
I L
aye
r
RS
S/a
tom
Co
mm
erc
e
Sto
rag
e
Se
curit
y
Me
ssa
gin
g
Ide
ntit
y
Ap
ps
La
yer
Ma
pp
ing
Ca
len
da
ring
Tra
nsa
ctio
n p
roc
Off
ice
pro
du
ctiv
ity
So
cia
l ne
two
rkin
g
Physical assets
Kn
ow
led
ge
La
yer
Do
ma
in/c
on
text
/SL
A
Dyn
am
ic a
lloca
tion
Co
llab
ora
tion
Exp
ert
ise
loca
tion
Co
ord
ina
tion
Me
asu
rem
en
t/m
on
itorin
g
Se
nse
ma
kin
g
Env
ironm
ent
SensoryLayer
Managing capabilities
Page:23 © 2006 Bala Iyer (bala@bu.edu)
Core capabilities
Fast searchContinuous innovation (beta products)Personalized results thru data integrationScalable platformCombinatorial innovation API/meta-data management
Page:24 © 2006 Bala Iyer (bala@bu.edu)
100,000 servers,Wi-Fi
GoogleOS
Search, Maps,Desktop
Applicationprovisioning
Google Enterprisesolutions
Systems(OS, file sharing, terminal access, db access, client-server)
Middleware Services(systems management, application & transaction servers, collaboration & messaging)
Application software(CRM package, SCM package, BI, HR, engineering & design, e-biz)
Services(IT consulting, systems integration, outsourcing, training & education,
maintenance)
Data management(meta data, distribution, schema evolution, temporal)
Hardware
API toolkit
Page:25 © 2006 Bala Iyer (bala@bu.edu)
Business principles
Defendable assets Do it yourself Presence in every layer of the industry stack Customer behavior data
Architecture wins business platform wars Manage emergent architecture Meta-data management
Ecosystem based innovation Modular development Services platform (APIs) Network based innovation
Two-sided marketsServing the Long tailTalent magnate
70/20/10Prediction markets
Page:26 © 2006 Bala Iyer (bala@bu.edu)
Multi-sided markets
Platform
Consumers (80 million)
E-business firms• Fortune 500 firms
Software developers• ISVs• Open source community
Deliver apps Support development
Core benefits
Page:27 © 2006 Bala Iyer (bala@bu.edu)
Wired issue 12.10 October 2004
Unique applications
Page:28 © 2006 Bala Iyer (bala@bu.edu)
70/20/10 principle
Source: Google Factory Tour
Page:29 © 2006 Bala Iyer (bala@bu.edu)
How does it impact your business model?1. Value proposition or value cluster2. Marketspace offering – product, service, information, or all three3. Defendable resource systems4. Revenue model
Page:30 © 2006 Bala Iyer (bala@bu.edu)
Resources
Great Giveaway by Erick Schonfeld, Business 2.0 April 2005.What’s next for Google by Charles Ferguson, TechnologyReview.com January 2005.Amazon: Giving away the store by Wade Roush, TechnologyReview.com January 2005.All the links cited in del.icio.us/balaiyerwww.google.com
Page:31 © 2006 Bala Iyer (bala@bu.edu)
Related Documents
