Structured concurrency control in object oriented databases

              STRUCTURED  CONCURRENCY  CONTROL     IN     OBJECT  ORIENTED  DATABASES                                     Francisco  Mariátegui                           STRUCTURED  CONCURRENCY  CONTROL   IN   OBJECT  ORIENTED  DATABASES             A  Dissertation  Presented  to  the  Graduate  Faculty  of   The  School  of  Engineering  and  Applied  Science   of   Southern  Methodist  University   in   Partial  Fulfillment  of  the  Requirements   for  the  Degree  of   Doctor  of  Philosophy   with  a   Major  in  Computer  Science   By   Francisco  José  Mariátegui   B.Sc.,  Honors,  Naval  Academy  of  Peru,  1974   Systems  Engineering  Specialization  Degree,  Honors,  University  of   Lima,  1977   M.Sc  Computer  Science,  U.S.A  Naval  Postgraduate  School,  1979   M.Sc  Computer  Systems  Management  U.S.A  Naval  Postgraduate   School,  1979         May  13,  1989                               COPYRIGHT  @  1989     Francisco  J  Mariategui     All  Rights  Reserved                       Mariategui,  Francisco  J       B.Sc  Naval  Sciences,  Naval  Academy  of  Peru,  1974   System  Engineering,  University  of  Lima,  1977   M.Sc  Computer  Science,  U.S  Naval  Postgraduate  School,  1979   M.Sc  Computer  Systems  Management,  U.S  Naval  Postgraduate  School,  1979       STRUCTURED  CONCURRENCY  CONTROL  IN     OBJECT  ORIENTED  DATABASES       Advisor:  Dr  Margaret  H  Eich     Doctor  of  Philosophy  degree  conferred  August  12,  1989     Dissertation  completed  May  13,  1989     In   the   last   few   years   a   number   of   object-­‐oriented   database   systems   have   appeared   in   the   literature,   most   of   which   addresses   specific   areas   such   as   office   information   systems   (OIS),   computer   aided   design   (CAD),   computer   aided   manufacturing   (CAM),   software   engineering   (SE),   and   artificial   intelligence   (AI)   Unfortunately,   hardly   any   one   of   them  addresses  the  problem  of  concurrency   control  from  the  general-­‐ purpose   database   point   of   view   Due   to   the   extreme   differences   in   types  of  transactions  supported  by  these  environments,  the  need  for               combining   different   concurrency   control   approaches   has   been   recognized  but  never  thoroughly  investigated     A  high  level  design  of  a  Multi-­‐Group  Multi-­‐Layer  approach  to   concurrency  control  for  object-­‐oriented  message-­‐passing  based   databases  is  presented  The  design  follows  a  formal  definition  of   transaction  The  concurrency  control  takes  advantage  of  the   structured  nature  of  transactions  to  manage  an  on-­‐line  serializer  The   serializer  is  specified  as  a  set  of  filters  These  filters  are  specifications   of  algorithms  that  ensure  serializable  histories  The  concurrency   control  manages  these  histories  by  layers  Each  layer,  along  with  its   corresponding  filters,  constitutes  a  different  level  of  abstraction  in   concurrency  control  processing  Mutually  exclusive  groups  of   transactions  being  processed  in  parallel  are  assumed  The  availability   of  a  processor  per  group  is  also  assumed  The  performance  is   improved  when  this  case  of  large  granularity  and  limited  interaction  is   applied  The  decomposition  of  the  histories  into  layers  allows  the   problem  to  be  more  manageable,  the  principles  of  hierarchical  design   to  be  applied,  and  the  benefits  of  hierarchical  thought  to  be  utilized     Summarizing,  this  research  has  led  to  the  following  results:                 1) First   cut   definition   of   an   Object-­‐Oriented   Data   Model   (OODM)   which   encompasses   data   structures,   operations,   and   integrity   constraints   2) Transaction   processing   model   for   the   OODM   environment,   which   facilitates   not   only   definition   of   transactions   but   also,   allows   investigation  of  concurrency  control   3) Multi-­‐group   Multi-­‐Layer   concurrency   control   technique   built   on   the   OODM   and   transaction   models   that   allow   the   use   of   several   different   concurrency   control   techniques   in   parallel   in   the   same   environment                   TABLE  OF  CONTENTS   TABLE  OF  CONTENTS     LIST  OF  FIGURES   13   ACKNOWLEDGEMENTS   15   CHAPTER  1  -­‐  INTRODUCTION   16   1.1  THE  PROBLEM   1.2  THE  APPROACH   1.3  CONTRIBUTION   1.4  SIGNIFICANCE   1.5  THE  CONCURRENCY  CONTROL  MANAGER   1.5.1  PURPOSE   1.5.2  CONCEPTS  AND  MEANS   1.5.3  BENEFITS   1.5.4  INTERFACE   1.6  GENERAL  OVERVIEW  OF  THE  CCMM   1.7  OUTLINE  OF  THE  DISSERTATION   2.1  INTRODUCTION   2.2  OBJECT-­‐ORIENTED  DATABASES:  AN  OVERVIEW   2.2.1  BACKGROUND   2.2.2  DEFINITION  OF  TERMS   2.2.3  DEFINITION  OF  PROPERTIES  OF  OODBS   2.3  DATA  MODELS   2.4  AN  OBJECT-­‐ORIENTED  DATA  MODEL   2.4.1  DATA  STRUCTURE   16   18   19   20   21   21   22   22   23   24   28   30   32   33   35   38   41   44   45               2.4.2  OPERATORS   2.4.3  INTEGRITY  RULES   2.4.4  SUMMARY   2.5  MODELING  ABILITY  OF  THE  OODM   2.6  SUMMARY   57   62   63   65   67   CHAPTER  3  -­‐  UNIT  OF  CONSISTENCY   69   3.1  INTRODUCTION   3.2  PRELIMINARIES   3.4  THE  PAIR  <  GM,  DM  >  AS  A  MODELING  TOOL   3.5  EXPANDING  THE  MODEL  TO  INCLUDE  LEAVES   3.6  THE  COMPLETE  STRUCTURE  OF  A  TRANSACTION   3.5  SUMMARY   69   70   79   87   93   102   CHAPTER  4  -­‐  THEORY  OF  EXECUTION  AND  SERIALIZABILITY   104   4.1  INTRODUCTION   4.2  PRELIMINARIES   4.3  SERIALIZABILITY   4.4  SERIALIZABILITY  AND  THE  PAIR     4.5  INFORMAL  COMPRESSION  OF  TRANSACTION  TREES   4.6  FROM  TRANSACTION  TREES  TO  TRANSACTION  HISTORIES   4.7  SUMMARY   104   105   107   118   119   128   130   CHAPTER  5  -­‐  MULTI-­‐LAYER  CONCURRENCY:  RATIONALE   132   5.1  INTRODUCTION   5.2  THE  NEED  FOR  CHANGE   5.3  REPRESENTATIVE  CONCURRENCY  CONTROL  TECHNIQUES   132   133   137               5.4  STRUCTURED  CONCURRENCY  CONTROL   5.5  THE  LAYERED  APPROACH   5.6  EXPANDING  THE  THEORY  OF  EXECUTION:  GROUP  HISTORY   5.7  SUMMARY   142   144   145   147   CHAPTER  6  -­‐  MULTI-­‐LAYER  CONCURRENCY  ARCHITECTURE   149   6.1  INTRODUCTION   6.2  BACKGROUND   6.3  ACTIVE  HISTORIES   6.4  PREFIXES   6.5  CONTENTS  OF  PREFIXES  OF  ACTIVE  HISTORIES   6.5.1  NOTATION  FOR  PREFIXES   6.5.2  MEANING  OF  THE  INDEXES  OF  HISTORIES   6.5.3  CONTENTS  OF  HISTORIES   6.6  HIERARCHY  OF  HISTORIES   6.7  ELEVATOR  FUNCTIONS   6.8  NOTATION   6.9  FILTERS   6.9.1  COMPONENT  HISTORIES  FILTERS   6.9.1.1  Filter  F0rsws   6.9.1.2  Filter  F0ccg   6.9.2  TRANSACTION  HISTORIES  FILTERS   6.9.2.1  Filter  F1rsws   6.9.2.2  Filter  F1ccg   6.9.3  FOREST  HISTORIES  FILTERS   6.9.3.1  Filter  F2rsws   6.9.3.2  Filter  F2tcg   6.9.4  GROUP  HISTORY  FILTERS       149   150   151   155   158   158   159   160   164   165   173   175   177   179   181   182   183   184   185   186   187   189   10         8.2.5  Early  Evaluation  of  Inter-­‐Group  Conflicts   The   Multi-­‐Group   Multi-­‐Layer   Approach   to   concurrency   "waits"   until   conflicts   have   gotten   to   the   top   of   the   hierarchy   to   detect   them   and   take   appropriate   action   It   is   in   this   way   that   inconsistencies   among   groups   are   not   allowed   The   explicit   maintenance   of   the   conflict   graphs   constitutes   not   only   the   way   to   guarantee   serializable   executions,   but   it   is   also   the   common   denominator   to   maintain   all   the   different  techniques  synchronized   It  may  seem  (at  first  glance)  that  the  detection  of  conflicts  at  the  top   level   of   the   hierarchy   is   a   "little"   too   late   (actually,   it   is   not)   The   underlying   assumption   of   this   work   is   that   different   types   of   transactions,   or   applications,   or   environments   must   be   supported   in   the   same   (logical)   data   bank   Furthermore,   this   data   bank   ought   to   provide   flexible   and   adaptable   concurrency   control   techniques   so   each   user   would   be   able   to   suit   their   needs   (or   lack   of   thereof)   It   is   more   likely   than   not,   that   different   transactions   will   access   different   data  due  to  their  different  nature,  their  different  type  of  application,   and  most  important,  their  different  environment  In  other  words,  the   probability  of  conflicts  among  groups  should  be  low  But,  then  again,       261         every   concurrency   control   technique   works   better   when   the   conflict   probability  is  low   In  any  event,  and  as  the  title  of  this  subsection  suggests,  it  would  be   interesting  to  learn  about  the  potential  benefits  of  early  evaluation  of   conflicts  There  are  several  early  evaluation  times:  levels  0,  1,  and  2  of   the   hierarchy   (ℎ𝐻)   In   the   case   that   these   early   evaluations   were   beneficial,  the  issue  about  how  much  space  is  needed  would  become   a   problem   But   more   important,   global   data   structures   would   be   needed   This   would   make   it   problematic   to   implement   in   distributed   systems   8.2.6  Increase  the  Number  of  Processors   So   far,   the   existence   of   only   four   processors   was   assumed,   one   per   group,   performing   concurrency   control   processing   At   the   top   of   the   hierarchy  and  given  the  sequential  nature  of  its  filters,  any  processor   (out   of   the   four)   may   execute   The   processing   needed   to   be   performed  at  each  level  of  the  hierarchy  is  simple  It  consists  mainly  of   cycle   detections   and   updates   to   data   structures   containing   read   sets   and  write  sets  and  other  bookkeeping  information   The   potential   for   speedup   using   parallel   algorithms   and   an   appropriate   number   of   processors,   is   substantial   For   example,       262         Kruskal's   parallel   prefix   algorithm   [Kruskal,   1985],   which   could   be   interpreted   as   a   variant   of   the   precede   and   conflict   relations,   has   a   time  complexity  of   O(  (  (𝑙𝑜𝑔  𝑛)/𝑙𝑜𝑔(2𝑛/𝑝    )).(𝑛/𝑝  )),     where  𝑝  in  the  number  of  processors  and  𝑛  the  number  of  elements     The   creation   of   a   hardware   version   of   the   Concurrency   Control   Manager  Module  (CCMM)  is  possible  This  version  could  take  the  form   of   a   concurrency   control   board,   which   among   other   things   (such   as   memory)  would  contain  a  set  of  processors  acting  on  the  hierarchy  of   histories   The   speedup   could   be   equivalent   to   several   orders   of   magnitude   when   parallel   processors   are   used   instead   of   only   one   processor  per  group   8.2.7  Pipeline  the  Cycle  Algorithm   As  is,  the  cycle  algorithm  is  processed  a  step  at  a  time,  that  is,  a  level   of   the   hierarchy   at   a   time   An   improvement   over   the   original   high-­‐ level   design   could   be   achieved   by   pipelining   the   steps   of   the   cycle   algorithm       263         Given   that   the   algorithm   is   described   in   terms   of   elevator   functions,   filters,  and  bookkeeping  information,  it  would  be  relatively  simple  to   pipeline   the   processing   of   different   levels   of   the   hierarchy   One   would   need  to  investigate  which  would  be  the  collision  vector  of  the  whole   scheme   In   other   words,   what   are   the   resources   used   by   each   level   and   how     they   collide   Here,   resources   mean   logical   as   well   as   physical  structures  (e.g.,  memory,  cpu's,  read  sets,  etc.)  By  “collide”  it   is  meant  if  they  are  used  simultaneously,  and  if  they  are,  how  are  they   going  to  interfere  with  each  other   Once   this   "interference"   is   understood,   the   pipeline   mechanism   may   be   constructed   and   optimized,   giving   a   further   gain   in   performance   The   gains   in   using   pipelines   can   be   quite   impressive   As   a   reminder,   consider   that   the   so-­‐called   "supercomputers"   are   (in   their   majority)   highly  pipelined  vector  processing  architectures   8.2.8  Router  Issues   It   is   assumed   throughout   this   work   that   there   exists   a   router   that   would   balance   the   load   in   the   groups   The   router,   as   mentioned   in   chapter   7,   should   be   smart   enough   to   assign   each   transaction   to   its   "most   appropriate"   group   and   at   the   same   time   perform   load   balancing  These  two  minimum  requirements  could  be  conflicting  At       264         this   point   there   are   no   definitive   answers   to   put   forward   concerning   this  problem  although  there  are  some  suggestions     Basically,  the  router  may  be  qualified  by  its  level  of  achievement,  i.e.,   how   well   would   it     its   job   Also,   the   quality   of   the   router   may   be   directly  dependent  on  the  number  of  characteristics  of  transactions  it   is   able   to   analyze   If   complexity   were   directly   correlated   with   capability,   then   the   router   would   be   more   efficient   if   it   is   more   complex  (up  to  a  certain  point,  of  course)   A   router   could   be   a   simple   round   robin   algorithm   that   would   assign   the   first   transaction   to   the   first   group,   the   second   transaction   to   the   second   group,   and   so   on   Or   it   can   be   a   complex   analyzer   of   a   transaction  or  a  group  of  transactions  using  techniques  from  artificial   intelligence,   statistics,   optimization,   etc   The   definitive   answer   may   lie   somewhere   in   between   The   router   should   be   simple   enough   to     the   job   efficiently   and   complex   enough   to   satisfy   most   of   the   requirements   Another  unanswered  question  about  the  router  is  where  to  situate  it   In   other   words,   is   the   router   a   new   and   separate   module?   It   was   assumed   throughout   this   work   that   this   is   the   case,   for   purposes   of   modularizing   the   presentation   and   thought   It   could   be   seen   as   a       265         smart   transaction   manager   or   even   a   combination   of   a   query   processor  and  a  transaction  manager  driven  by  an  expert  system   The   two   questions   presented   above   must   be   answered   before   an   attempt  is  made  to  study  the  impact  that  such  a  device  will  have  when   included  in  the  DBMS             266         REFERENCES     [Aho,  1974]   Aho,   Alfred   V.,   Hopcroft,   John   E.,   and   Ullman,   Jeffrey   E.,   The   Design   and  Analysis  of  Computer  Algorithms,  Addison-­‐Wesley,  1974     [Aho,  1983]   Aho,   Alfred   V.,   Hopcroft,   John   E.,   and   Ullman,   Jeffrey   E.,   Data   Structures  and  Algorithms,  Addison-­‐Wesley,  1983     [Baase,  1978]   Baase,   S.,   Computer   Algorithms,   Introduction   to   Design   and   Analysis,   Addison-­‐Wesley,  1978     [Baase,  1988]   Baase,   S.,   Computer   Algorithms,   Introduction   to   Design   and   Analysis,   Second  Edition,  Addison-­‐Wesley,  1988     [Badrinath,  1987]   Badrinath   B   and   Ramamritham,   K.,   "Semantics-­‐based   Concurrency:   Beyond   Commutativity,"   Proceedings   of   the   Third   International   Conference  on  Data  Engineering,  1987,  pp  304-­‐311     [Badrinath,  1988]   Badrinath   B   and   Ramamritham,   K.,   "Synchronizing   Transactions   on   Objects,"   IEEE   Transactions   on   Computers,   Vol   37,   No   5,   May   1988,   pp  541-­‐547     [Baneriee,  1987]       267         Banerjee  J.,  Chou,  H.T.,  Garza,  J.F.,  Kim,  W.,  Woelk,  D.,  Ballou,  N.,  and   Kim,  H.J.,  "Data  Model  Issues  for  Object-­‐Oriented  Applications,"  ACM   Transactions   on   Office   Information   Systems,   Vol   5,   No   1,   January   1987,  pp  3-­‐26     [Beech,  1987]   Beech,   D.,   "Groundwork   for   an   Object   Database   Model,"   Research   Directions   in   Object-­‐Oriented   Programming,   MIT   Press   Series   in   Computer  Systems,  Cambridge,  Massachusetts,  1987,  pp  317-­‐355     [Bernstein,  1981]   Bernstein,   P   and   Goodman,   N.,   "Concurrency   Control   in   Distributed   Database  Systems,"  ACM  Computing  Surveys,  July  1981,  pp  185-­‐222     [Bernstein,  1987]   Bernstein,   P.,   Hadzilacos,   V.,   and   Goodman,   N.,   Concurrency   Control   and  Recovery  in  Database  Systems,  Addison-­‐Wesley,  1987     [Casanova,  1981]   Casanova,   M.A.,   The   Concurrency   Control   Problem   for   Database   Systems,  Lecture  Notes  in  Computer  Science,  Springer-­‐Verlag,  1981     [Cockshot,  1984]   Cockshot,   W.P.,   Atkinson,   M.P.,   Chrisholm,   J.,   Bailey,   P.J.,   and   Morrison,   R.,   "Persistent   Object   Management   System,"     Software-­‐Practice  and  Experience,  Vol  14,  1984,  pp  49-­‐71     [Codd,  1970]   Codd,   E.F.,   "A   Relational   Model   for   Large   Shared   Data   Banks,"   Communications  of  the  ACM,  Vol  13,  No  8,  June  1970,  pp  377-­‐387         268         [Codd,  1981]   Codd,   E.F.,   "Data   Models   in   Database   Management,"   ACM   SIGMOD   Record,  Vol  11,  No  2,  February  1981     [Codd,  1982]   Codd,   E.F.,   "Relational   Database:   A   Practical   Foundation   for   Productivity,"   Communications   of   the   ACM,   Vol   25,   No   2,   February   1982,  pp  109-­‐117     [Croft,  1985]   Croft,   W.B.,   "Task   Management   for   an   Intelligent   Interface,"  Database   Engineering,  Vol  8,  No  4,  December  1985,  pp  8-­‐13     [Dahl,  1966]   Dahl  and  Nygaard,  Communications  of  the  ACM,  Vol  9,  pp  671-­‐678     [Date,  1983]   Date,   C.J.,   An   Introduction   to   Database   Systems,   Volume   II,   The   Systems  Programming  Series,  Addison-­‐Wesley,  1984     [Date,  1986]   Date,   C.J.,   An   Introduction   to   Database   Systems,   Volume   1,   fourth   edition,  Addison-­‐Wesley  Publishing  Company,  1986     [Dayal,  1986]   Dayal,   U   and   Smith,   J.M.,   "PROBE:   A   Knowledge-­‐Oriented   Database   Management   System,"   On   Knowledge   Base   Management   Systems,   Springer-­‐Verlag,  1986,  pp  227-­‐257     [Eich,  1988]       269         Eich,   M.H.,   "Graph   Directed   Locking,"   IEEE   Transactions   on   Software   Engineering,  Vol  14,  No  2,  February  1988,  pp  133-­‐140     [ELAbbadi,  1988]     ElAbbadi,   M.,   "The   Group   Paradigm   for   Concurrency   Control   Protocols",   Proceedings   of   the   International   Conference   SIGMOD   on   Mgmt    of  Data,  Chicago,  Illinois,  June  1-­‐3  1988,  pp  126-­‐134     [Eswaran,  1976]   Eswaran,  K.P.,  Gray,  J.N.,  Lorie,  R.A.,  and  Traiger,  I.L.,  "The  Notions  of   Consistency   and   Predicate   Locks   in   a   Database   System,"   Comm   of   the   ACM,  Vol  19,  No  11,  November  1976,  pp  624-­‐633     [Fishman,  1987]   Fishman,  D.,  Beech,  D.,  Cate,  H.,  Chou,  E.,  Connors,  T.,  Davis,  J.,  Derret,   N.,  Hoch,  C.,  Kent,  W.,  Lynbaek,  P.,  Mahbod,  B.,  Neimat,  N.,  Ryan,  T.,   and   Shan,   M.,   "Iris:   An   Object-­‐Oriented   Database   Management   System,"  ACM  Transactions  on  Office  Information  Systems,  Vol  5,  No   1,  January  1987,  pp  48-­‐69     [Garcia-­‐Molina,  1987]   Garcia-­‐Molina,   H   and   Salem,   K.,   "Sagas,"   Proceedings   of   the   ACM   International   Conference   on   Management   of   Data,   May   1987,   pp   249-­‐259     [Garey  ,1979]   Garey,   M   and   Johnson,   D.,   Computers   and   Intractability,   A   Guide   to   the  Theory  of  NP-­‐Completeness,  W.H  Freeman  and  Company,  1979     [Goldberg,  1983]       270         Goldberg   A   and   Robson,   D.,   Smalltalk-­‐80:   The   language   and   its   implementation,  Reading,  Massachusetts,  Addison-­‐Wesley,  1983     [Gray,  1982]   Gray,   J.,   Lorle,   R.,   Putzolu,   G.,   and   Traiger,   I.,   "Granularity   of   Locking   and  Degrees  of  Consistency  in  a  Shared  Data  Base,"  Modeling  in  Data   Base  Management  Systems,  G.M  Nijssen,  ed.,  North  Holland,  1976     [Hadzilacos  ,1986]   Hadzilacos,  T  and  Yannakakis,  M.,  "Deleting  Completed  Transactions,"   Proceedings   of   the   Fifth   ACM   SIGACT-­‐SIGMOD   Symposium   on   Principles  of  Data  Base  Systems,  pp  43-­‐46,  1986     [Harary,  1969]   Harary,  E.,  Graph  Theory,  Addison-­‐Wesley,  1969     [Horrowitz,  1978]   Horrowitz,   E   and   Sahni,   S.,   Fundamentals   of   Computer   Algorithms,   Computer  Science  Press,  1978,  pp  248-­‐269     [Knuth,  1975]   Knuth,   D.,   Fundamental   Algorithms:   The   Art   of   Computer   Programming,  Volume  1,  Addison-­‐Wesley,  Second  Edition,  1975     [Kohler  ,1981]   Kohler,  W.,  "Survey  of  Techniques  for  Synchronization  and  Recovery  in   Decentralized   Computer   Systems,"   ACM   Computing   Surveys,   June   1981,  pp  149-­‐183     [Kruskal,  1985]       271         Kruskal,   C.,   Rudolph,   L.,   and   Snir,   M.,   "The   Power   of   Parallel   Prefix,"   IEEE  Transactions  on  Computers,  Vol  c-­‐34,  No  10,  1985     [Kung,  1981]   Kung,   H   and   Robinson,   J.,   "On   Optimistic   Methods   for   Concurrency   Control,"  ACM  Transactions  on  Database  Systems,  Vol  6,  No  2  June   1981,  pp  213-­‐226     [Lieberman,  1981]   Lieberman,   H.,   "A   Preview   of   Act   1,"   MIT   Artificial   Laboratory   Memo   No  625     [Lochovsky,  1985]   Lochovsky,  F.,  "Letter  from  the  editor,"  Database  Engineering,  Vol  8,   No  4,  December  1985,  page  1     [Maier,  1983]   Maier,  D.,  The  Theory  of  Relational  Database,  Computer  Science  Press,   1983     [Maier,  1986]   Maier,   D   and   Stein,   J.,   "Development   of   an   Object-­‐Oriented   DBMS,"   OOPSLA  Proceedings,  1986,  pp  472-­‐482     [Maier,  1988]   Maier,  D.,  Penney,  J.,  and  Stein,  J.,  "Is  the  Disk  Half  Full  or  Half  Empty,"   To  appear:  Workshop  on  Persistent  Object  Systems,  Appin,  Scotland     [Manthey,  1987]   Manthey,   J.,   The   Characteristics   of   Parallel   Algorithms,   Scientific   Computation  Series,  MIT  Press,  1987,  pp  139-­‐165       272           [Mariategui,  1988]   Mariategui,   F.,   Eich,   M.,   and   Rafiqi,   S.,   "The   Object-­‐Oriented   Data   Model  Defined,"  March  1988  SMU  Technical  Report  88-­‐CSE-­‐28     [McKenzie,  1987]   McKenzie,   E   and   Snodgrass,   R.,   "Extending   the   Relational   Algebra   to   Support   Transaction   Time,"   Proceedings   ACM   SIGMOD   Annual   Conference,  San  Francisco,  May  27-­‐29,  1987,  pp  467-­‐478     [Moss,  1981]   Moss,   J.,   "Nested   Transactions:   An   Approach   to   Reliable   Distributed   Computing",   Ph.D   Thesis   M.I.T   Dept   of   Elec   Eng   and   Comp   Sci.,   1981     [Moss,  1985]   Moss,   J.,   Nested   Transactions:   An   Approach   to   Reliable   Distributed   Computing,  The  MIT  Press  Series  in  Information  Systems,  1985     [Ontologic,  1986]   Vbase  Technical  Overview,  Version  1.0,  March  6,  1987,  1986  Ontologic   Inc     [Papadimitriou,  1979]   Papadimitriou,   C.,   "The   Serializability   of   Concurrent   Database   Updates,"  Journal  of  the  Association  for  Computing  Machinery,  Vol  26,   No.4,  October  1979,  pp  631-­‐653     [Papadimitriou,  1986]   Papadimitriou,   C.,   The,   Theory   of   Database,   Concurrency   Control,   Computer  Science  Press,  1986       273           [Peterson,  1985]   Peterson   J   and   Silberschatz,   A.,   Operating   System   Concepts,   second   edition,  Addison-­‐Wesley  Publishing  Company,  1985     [Serviologic,  1988]   Servio  Logic  Development  Corporation,  Beaverton,  Oregon,  1988     [Segev,  1987]   Segev   A   and   Shoshani,   A.,   "Logical   Modeling   of   Temporal   Data",   Proceedings   of   ACM  SIGMOD   Annual   Conference,  San  Francisco,  May   27-­‐29,  1987,  pp  454-­‐466     [Shapiro,  1977]   Shapiro,   R.M   and   Millstein,   R.E.,   NSW   Reliability   Plan,   Technical   Report  7701-­‐1411,  Computer  Associates,  Wakefield,  MA,  June,  1977     [Shmueli,  1983]   Shmueli,   O.,   "Dynamic   Cycle   Detection,"   Information   Processing   Letters,  17:4,  1983     [Silberschatz,  1982]   Silberschatz,   A   and   Kedem,   Z.,   "A   Family   of   Locking   Protocols   for   Database   Systems   that   are   Modeled   by   Directed   Graphs,"   IEEE   Transactions  on  Software  Engineering,  Vol  SE-­‐8,  No  6,  Nov  1982     [Stefik,  1986]   Stefik,   M   and   Bobrow,   D.,   "Object-­‐Oriented   programming:   Themes   and  Variations,"  AI  Magazine,  January  1986,  pp  40-­‐62     [Ullman,  1982]       274         Ullman,  J.,  Principles  of  Database  Systems,  second  edition,  Computer   Science  Press,  1982     [Ullman,  1988]   Ullman,   J.,   Principles   of   Database   and   Knowledge-­‐Base   Systems,   Volume  1,  Computer  Science  Press,  1988     [Weinreb  1981]   Weinreb,  D  and  Moon  D.,  "Lisp  Machine  Manual,"  Symbolics  Inc     [Woelk,  1986]   Woelk,  D.,  Kim,  W.,  and  Luther,  W.,  "An  Object-­‐Oriented  Approach  to   Multimedia  Databases,"  Proceedings  ACM  SIGMOD  Conference  on  the   Management  of  Data,  Washington  D.C.,  May  1986     [Zaniolo,  1986]   Zaniolo,   C.,   Hassan   A.,   and   Beech   D.,   Cammarata   Stephanie,   Kerschberg   Larry,   and   Maier   David,   "Object-­‐Oriented   Database   Systems   and   Knowledge   Systems,"   Proceedings   1st   International   Workshop  on  Expert  Database  Systems,  1986,  pp  293-­‐305     [Zhu,  1987]   Zhu,   J   and   Maier,   D.,   "Abstract   Object   in   an   Object-­‐Oriented   Data   Model,"  Oregon  Graduate  Center  Department  of  C.S  and  Engineering   Technical  Report    No  CS/E  87-­‐015           275   ...         STRUCTURED CONCURRENCY CONTROL   IN   OBJECT ORIENTED DATABASES             A  Dissertation  Presented  to  the  Graduate  Faculty  of   The  School  of  Engineering  and  Applied...  to   concurrency control  for object- ­ oriented  message-­‐passing  based   databases  is  presented  The  design  follows  a  formal  definition  of   transaction  The concurrency control. .. databases   in   general,   and   object- ­ oriented   databases   in   particular   An   approach   to   a   solution   can   be   accomplished   by   focusing   our   efforts   in   Structured   Concurrency