BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:A comparative survey of locking in ordered indexes
DTSTART:20160923T120000
DTEND:20160923T133000
DTSTAMP:20260929T061546Z
UID:8d024a654eca1db18d4e1dd6c3b04dbf23d5a906a9660fe5014cbcfc
CATEGORIES:Conferences - Seminars
DESCRIPTION:Goetz Graefe\, Google\nFor decades\, b-tree data structures ha
 ve been ubiquitous in databases\, file systems\, and information retrieval
 \; and over the past 25 years\, record-level locking has become ubiquitous
  for ordered indexes such as b-trees. There are multiple designs for fine-
 granularity locking in b- tree indexes\, each a different tradeoff between
  (i) a fine granularity of locking for high concurrency during updates\, (
 ii) coarse locks for equality queries and range queries\, (iii) run-time e
 fficiency with the fewest possible invocations of the lock manager\, and (
 iv) conceptual simplicity for efficient development\, maintenance\, and te
 sting. Using specific examples such as insertions\, deletions\, equality q
 ueries\, and phantom protection\, this case study compares five alternativ
 e designs for fine-granularity locking in b-tree indexes. For queries\, on
 e design dominates all prior other including all designs in industrial use
  today. For updates\, the same design is practically equal to the most rec
 ent prior design\, which dominates all other ones. These results\, togethe
 r with experiments reported in the past\, suggest that new b-tree implemen
 tations as well as existing ones ought to adopt a new locking technique.Fo
 r decades\, b-tree data structures have been ubiquitous in databases\, fil
 e systems\, and information retrieval\; and over the past 25 years\, recor
 d-level locking has become ubiquitous for ordered indexes such as b-trees.
  There are multiple designs for fine-granularity locking in b- tree indexe
 s\, each a different tradeoff between (i) a fine granularity of locking fo
 r high concurrency during updates\, (ii) coarse locks for equality queries
  and range queries\, (iii) run-time efficiency with the fewest possible in
 vocations of the lock manager\, and (iv) conceptual simplicity for efficie
 nt development\, maintenance\, and testing. Using specific examples such a
 s insertions\, deletions\, equality queries\, and phantom protection\, thi
 s case study compares five alternative designs for fine-granularity lockin
 g in b-tree indexes. For queries\, one design dominates all prior other in
 cluding all designs in industrial use today. For updates\, the same design
  is practically equal to the most recent prior design\, which dominates al
 l other ones. These results\, together with experiments reported in the pa
 st\, suggest that new b-tree implementations as well as existing ones ough
 t to adopt a new locking technique.
LOCATION:BC 410 https://plan.epfl.ch/?room==BC%20410
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR
