Re: se vs online engines
Naresh Chainani wrote: > Here are some differences between SE & online : Actually, these are mainly differences between OnLine 5.x and other versions of OnLine (OnLine-DSA, ODS, IDS, IUS, XPS, Foundation, you name it). It is courteous to give attributions for such work, too. > What are the fundamental limitations of OnLine 5.x compared with other > databases (especially later Informix databases such as 7.x, 8.x, 9.x, > but also Oracle, DB2, etc)? Here's a list in no particular sequence, > created with some help from: ...and a list of names was excised... This is a document that I assembled and posted to an internal mail alias. This version is missing a few OnLine 5.x limitations and features: * 1000 users maximum (including administrative processes) * Limited number of logical logs (61?) * No SMI; no sysmaster database * Parallel sort (in general; not just indexes) Note that the x in 5.x means x = 10 for most practical purposes. Triggers were not in 5.00 but were in 5.01, for example. > ------------------------------------------------------------------------ > > * 2 GB chunk sizes. > * Only TB-Tape as a backup system. > * Limited recovery options (no point-in-time recovery, no dbspace > level backup/restore; physical only restores). > * One sqlturbo process per user (no multithreading engine). > * I-Star (networking) is a separate product. > * ESQL/C 5.x has no support for connections. > * Limited SQL-92 (let alone SQL-99) features. > * No GLS (though some might regard that as a virtue!). > * No features for Oracle support. > * No replication (neither HDR nor ER). > * No extensibility. > * No parallel query. > * No High Performance Loader. > * No MaxConnect. > * No fragmentation (which limits the size of a table). > * No RAS features. > * Limited tbstat capabilities. > * Lack of tools for support/diagnosis (xtrace, dbcokpit, onperf, etc). > * No temp dbspaces. > * No support for (XPS) raw tables. > * Cannot change size of logical logs or physical log without > re-initializing instance. > * Limited JDBC support (via CSDK and I-Star/I-Net rather than ESQL/C > 5.x). > * Limited ODBC support (via CSDK and I-Star/I-Net rather than ESQL/C > 5.x). > * No data distributions for optimizer (but simple UPDATE STATISTICS). > * Small limit on number of of chunks due to non expandable reserved > pages, hence small total database size. > * No ISA support (Informix Server Administration). > * No external backup restore (EBR). > * No fuzzy checkpoints. > * Lock contention problems (adjacent key locking) on insert and delete > statements. > * Turbo processes that would not die after remote client dies. > > Features OnLine 5.x does have: > * XA transactions > * Triggers > * Parallel index build (PSORT_NPROCS, PSORT_DBTEMP) > * Simplicity > * Nice and stable database * Stored procedures * Small disk footprint * Smaller memory footprint for small systems * DTR (discrete triggered replication?) was available if you knew who to ask and had enough money to pay for it. > ------------------------------------------------------------------------ > > Mary Biggs wrote: > > We have a database program using Informix Standard Engine 7.3 running on > > SCO Unix. Most of the documentation I see refers to Informix Online > > engines. Is SE managed differently from Online? If so, where is the > > best source of information for learning to manage/maintain the SE engine > > database? Now, what I don't have handy is a summary of the differences between SE and the OnLine family of engines. -- Yours, Jonathan Leffler (Jonathan.Leffler@Informix.com) #include <disclaimer.h> Guardian of DBD::Informix v1.00.PC1 -- http://www.perl.com/CPAN "I don't suffer from insanity; I enjoy every minute of it!"
On Thu, 19 Oct 2000 18:19:41 +0100, Jonathan Leffler <jleffler@informix.com> wrote: <snip> > >Now, what I don't have handy is a summary of the differences between SE >and the OnLine family of engines. SE has (speaking as a Unix user): 1) No concept of different instances, just different databases. 2) All tables exist as two files (.dat & .idx) in the unix file system - raw disk is not possible 3) Less performance 4) Much easier to set up & administer 5) Can't do backups whilst people are using the database, or at least not snapshots. 6) Holds locks in Unix lock table (do any still use the other way) so if a lock gets left over the machine has to be rebooted 7) If the machine falls over the administrator has to sort the database out - no automatic rollback of uncompleted transactions 8) changes to the database schema don't get logged in the transaction log 9) lots more I can't remember just now! > >-- >Yours, >Jonathan Leffler (Jonathan.Leffler@Informix.com) #include <disclaimer.h> >Guardian of DBD::Informix v1.00.PC1 -- http://www.perl.com/CPAN > "I don't suffer from insanity; I enjoy every minute of it!"