and sql/mm part 7: history
DESCRIPTION
32N1766. and SQL/MM Part 7: History. ISO/IEC JTC 1/SC 32 WG 4 SQL/MM Convener Kohji SHIBANO. Revenue Structure. Usual Enterprise. Google. Net income. Net income. 5%. $5B 25%. Equipment cost. Operating income $10B 50%. Operating income 20%. Labor cost. Equipment cost. - PowerPoint PPT PresentationTRANSCRIPT
and SQL/MM Part 7: History
ISO/IEC JTC 1/SC 32 WG 4 SQL/MM ConvenerKohji SHIBANO
32N1766
Revenue Structure
Procurement cost
Labor cost
Equipment cost
Net incom
eOperating income20%
Sales amount100%
Usual Enterprise
5%
Procurement cost
Labor cost
Equipment cost
Net incom
eOperating income$10B50%
Sales amont$20B100%
$5B25%
Google’s Businesses and Services
Business AdWord AdSense
Service Search
Web search Earth Map
Communicate, show & share Document Gmail YouTube
mobile
Google business model From Portal to AdWord
Google data processingGoogle’s PageRank was a technology breakthroughCrawling and PageRank computation requires a lot of computations
Thus Google develop a set of new technologies for their infrastructure
CrawlerText
Extraction
PageRank
Search results
Google servers
Cloud ComputingGoogle computational infrastructure
1 million PC20 PB/Day
Google File System ( GFS)Google Work
Queue ( GWQ)
Bigtable
MapReduce
Chubby (lock mgr)
Operating System
Database
Application Framework
Application Programming Interface
Responding search requests worldwide
Google Bigtable Data Model
(row:string, column:string, time:int64) → string
Google Bigtable applications
SQL/MM Approach Using SQL as a formal specification language
In late 70’s and early 80’s within IBM Research Criticized to use a formal method such as VDM (Vienna
Development Method) and VDL (Vienna Development Language) developed by IBM Vienna Lab for the specification of SQL language
In SC 21 (OSI), strong recommendation to use formal methods
SQL/MM adopt SQL as a formal specification language MM implementations includes
DB2, Oracle, PostGreSQL, MySQL etc. MM services are implemented directly
Performance optimizations are up to the implementers
SQL Part 7: History In early 90’s, Temporal Database
Inspired by temporal logic base In the 21st Century, computing environment
drastically changed Massive computational power and storage
capacity make things possible Massive computation Massive information storage including historical records
Thus history support in SQL is SQL/MM Part 7: History
SQL/MM requirements The current SQL functionalities can support
most of the functionalities found in Google’s cloud computing
Only lacked functionality is the support of “HISTORY”
Google Bigtable (row:string, column:string, time:int64) → string
SQL/MM Part 7: History