Full Disclosure Archive

  The following index was added to the schema to improve query performance
  
  CREATE INDEX M_lo_idx on M_largeorder(lo_id)

  The load programs were not modified

  No changes were made to the reference beans

  All beans were deployed using Container Managed Persistence (CMP)

  The database was configured with the default isolation-level of READ_COMMITED
  
  The following beans were deployed with an isolation level of 'READ_COMMITED'
  With this isolation-level, the appserver will load the entities at the beginning
  of the transaction and commit at the end of the transaction, if the entities have
  been modified during the course of the transaction
  
  RuleEnt
  CustomerEnt
  LargeOrderEnt
  SupplierEnt
  
  The following beans were deployed with an isolation level of 'OPTIMISTIC_REPEATABLE_READ'
  With this isolation-level, the appserver will load the entity once and cache it. At the
  end of the transaction, if the entity has changed a verified update (all fields) will occur.
  If successful, the cache is updated with the latest values. If the update was not successful
  the transaction is rolled back and cached entity removed. If the bean state has not changed,
  a verified select will ensure the entity has not changed in the database. If the verified
  select is not successful, the transaction is rolled back and cached entity is removed.
  
  DiscountEnt
  ItemEnt
  ComponentEnt
  AssemblyEnt
  BomEnt
  PartEnt
  SupplierCompEnt
  POEnt
  
  The following beans were deployed with the isolation level 'REPEATABLE_READ'
  With this isolation-level, the appserver will obtain a lock on the entity in
  the database (SELECT FOR UPDATE) at the begining of the transaction. If the entity
  has changed during the course of the transaction the changed values are commited to the database.
  
  SequenceEnt
  OrderLineEnt
  OrderCustomerEnt
  OrderEnt
  WorkOrderEnt
  InventoryEnt
  SComponentEnt
  POLineEnt

  Software RAID Level 1 disk mirroring was used to ensure
  durability of the disk used for database logs.

  Database data and log devices are created using reference scripts.
  The data device size and log device size are 18 GB each, which are
  sufficient for published transaction rate. Three disks were used
  1 for data and 2 for logs All the tables were created on a single disk.
  
  A 45 minute run at an injection rate of 34 increased storage use by
  0.3 GB. Given linear scaling, eight hours of run time would at that
  rate require 3.2 GB. The system was configured with 18GB of disk
  storage which is sufficient for the 8 hour run.

  Pramati Application Server uses pass-by-value semantics by default

  The xerces.jar file included in the SPECjAppServer2001 kit was added
  to the classpath and used. The version is not known.
  
  No software or hardware was used to influence network traffic.
  
  Load Balancing was not performed.

The ords.err and plannedlines.err contain ConnectExceptions. Due to momentary
spikes in the network load the RMI runtime received some 'connection refused
exceptions'.