google bigtable - ict.ac.cnprof.ict.ac.cn/dcomputing/uploads/2011/dc_6_bigtable.… · ·...
TRANSCRIPT
Google BigtableFay Chang, Jeffrey Dean, Sanjay Ghemawat, Wilson C. Hsieh, Deborah A. Wallach, Mike Burrows, Tushar Chandra, Andrew Fikes,
Robert E. GruberGoogle, Inc.
UWCS OS Seminar DiscussionErik Paulson
2 October 2006
See also the (other)UW presentation by Jeff Dean in September of 2005 (See the link on the seminar page, or just google for “google bigtable”)
2 of 19
Before we begin…
• Intersection of databases and distributed systems
• Will try to explain (or at least warn) when we hit a patch of database
• Remember this is a discussion!
3 of 19
Google Scale• Lots of data
– Copies of the web, satellite data, user data, email and USENET, Subversion backing store
• Many incoming requests• No commercial system big enough
– Couldn’t afford it if there was one– Might not have made appropriate design choices
• Firm believers in the End-to-End argument• 450,000 machines (NYTimes estimate, June 14th
2006
4 of 19
Building Blocks• Scheduler (Google WorkQueue)• Google Filesystem• Chubby Lock service• Two other pieces helpful but not required
– Sawzall– MapReduce (despite what the Internet says)
• BigTable: build a more application-friendly storage service using these parts
5 of 19
Google File System
• Large-scale distributed “filesystem”• Master: responsible for metadata• Chunk servers: responsible for reading
and writing large chunks of data• Chunks replicated on 3 machines, master
responsible for ensuring replicas exist• OSDI ’04 Paper
6 of 19
Chubby
• {lock/file/name} service• Coarse-grained locks, can store small
amount of data in a lock• 5 replicas, need a majority vote to be
active• Also an OSDI ’06 Paper
7 of 19
Data model: a big map•<Row, Column, Timestamp> triple for key - lookup, insert, and delete API
•Arbitrary “columns” on a row-by-row basis
•Column family:qualifier. Family is heavyweight, qualifier lightweight
•Column-oriented physical store- rows are sparse!
•Does not support a relational model
•No table-wide integrity constraints
•No multirow transactions
8 of 19
SSTable• Immutable, sorted file of key-value
pairs• Chunks of data plus an index
– Index is of block ranges, not values
Index
64K block
64K block
64K block
SSTable
9 of 19
Tablet
• Contains some range of rows of the table• Built out of multiple SSTables
Index
64K block
64K block
64K block
SSTable
Index
64K block
64K block
64K block
SSTable
Tablet Start:aardvark End:apple
10 of 19
Table• Multiple tablets make up the table• SSTables can be shared• Tablets do not overlap, SSTables can overlap
SSTable SSTable SSTable SSTable
Tabletaardvark apple
Tabletapple_two_E boat
11 of 19
Finding a tablet
12 of 19
Servers• Tablet servers manage tablets, multiple tablets
per server. Each tablet is 100-200 megs– Each tablet lives at only one server– Tablet server splits tablets that get too big
• Master responsible for load balancing and fault tolerance– Use Chubby to monitor health of tablet servers,
restart failed servers– GFS replicates data. Prefer to start tablet server on
same machine that the data is already at
13 of 19
Editing a table• Mutations are logged, then applied to
an in-memory version• Logfile stored in GFS
SSTable SSTable
Tablet
apple_two_E boat
Insert
InsertDelete
InsertDelete
Insert
Memtable
14 of 19
Compactions• Minor compaction – convert the memtable into
an SSTable– Reduce memory usage – Reduce log traffic on restart
• Merging compaction– Reduce number of SSTables– Good place to apply policy “keep only N versions”
• Major compaction– Merging compaction that results in only one SSTable– No deletion records, only live data
15 of 19
Locality Groups
• Group column families together into an SSTable– Avoid mingling data, ie page contents and
page metadata– Can keep some groups all in memory
• Can compress locality groups• Bloom Filters on locality groups – avoid
searching SSTable
16 of 19
Microbenchmarks
17 of 19
18 of 19
Application at Google
19 of 19
Lessons learned
• Interesting point- only implement some of the requirements, since the last is probably not needed
• Many types of failure possible• Big systems need proper systems-level
monitoring• Value simple designs