Conference ulysse::rdb_vms_competition

    The features/functions grid has been updated. The new one is available
    in the same location as the old one.
    (NOVA::PM01:[booth.public]RDBMS_features_grid.ps)
    
    I have never seen a direct independent comparison of Oracle/Rdb.
    
    ---- Michael Booth
    
    P.S.
    
    A "rave review" of any product is generally unacceptable as usually it
    means the writer has a strong bias. A mildly positive review can be
    just as heavily biased, but is much more credible.
    
    ---- Michael Booth

    Recently, a customer asked us here in London more or less the same 
    thing. They wanted to choose a database and didnt know how to go
    about it. They asked us to design a questionnaire which they could
    submit to various database vendors, including us of course!
    So, I put in this questionnaire things that I know Oracle is very
    weak at and Rdb is very strong at. The end result was that Oracle
    was defeated. Indeed, they could only answer some of the questions
    by referring to their version 6 which was not yet out at that time.
    
    Hope this helps.
    
	Johannes van Vuren.
	




				Page 1

	QUESTIONS ON ADVANCED DATABASE MANAGEMENT SYSTEM TECHNOLOGY
	-----------------------------------------------------------






	The questions are devided into the following catagories:

	A. Architecture
	B. Physical data management
	C. Logical data management
	D. Security, recovery and disaster management
	E. Clusters and networking
	F. Tuning and Performance
	G. Tools for end users, application designers and database
	   administrators
	H. Data dictionaries
	I. User manuals and technical information



	(Questions marked with an * are especially important)
		
				Page 2
	A. Architecture
	   ------------
	1. Give an overview of your database management system architecture.
       *2. Describe the view this architecture presents to end-users and
	   application designers.
	3. Describe to what extent the architecture is open or closed. Can the
	   database interface to third party products, especially to third
	   party data dictionaries and productivity tools.
	4. To what extent is your system bundled or un-bundled. Can its 
	   different parts be bought seperately. Describe the functional
	   dependencies between the different components.
	5. Does your system supports the 3-schema model.


	B. Physical data management
	   ------------------------
	1. Give an overview of your physical implementation of the Relational
	   Model and which performance issues are addressed by this
	   implementation.
       *2. Describe the controls available in the system to allow for the 
	   placement of data in multiple files and disk drives. Is your 
	   system a single or multi-file system.
	3. Can your database be dynamically restructured and can space
	   allocation and physical placement of data be altered while users
	   are using the system.
	4. Describe how your system handles physical space management,
	   especially how space is re-used on deletions. Describe how free
	   space is allocated and controlled.
       *5. Describe how the system uses fast access techniques eg. indexing,
	   clustering, hashing, etc.
	6. Describe how your system handles locking between user processes.
	   To what extent does it uses the lock manager of the underlying host
	   operating system. Is the locking your system uses adjustable.
	7. Describe the main advantages derived from the operation of the
	   Query Optimiser of your system. To what extent does the Optimiser
	   make use of indices, system relations and internal statistics, if
	   at all. To what extent can application designers enhance or by-pass
	   the Optimiser.
	8. Describe whether your system supports fixed and/or variable page
	   sizes. Can page sizes be adjusted to match data requirements.
	9. Describe the processes used by your system to read and write to 
	   the database. How do these processes handle grouping of I/O's in
	   order to maximise disk drive operation, especially in situations
	   where multiple drives are employed.
       *10.Describe whether your system supports a global cache of database 
	   pages and/or a per-user cache scheme. What are the advantages of
	   the method your system uses in terms of I/O activity and memory
	   use in a scenario where different users accesses different portions
	   of the database.
	11.How much manual intervention is necessary to ensure your system
	   uses space efficiently.
	12.Describe on which hardware your system has been best implimented
	   so far.
       *13.Can your system access data held in third party databases.
	   Please elaborate.
       *14.Does your system support Horizontal Partitioning.
	15.What facilities or additional products are available in your 
	   system to support the distribution of data to remote sites.








				Page 3

	C. Logical data management
	   -----------------------
	1. Give an overview of your logical implementation of the Relational
	   Model.
       *2. Describe which of the 12 relational rules (the Codd rules) and the 
	   additional rules your system complies with.
        3. Describe the relational operators your system supports.
       *4. To what extent does your system supports referential integrity and
	   how is this support visible to your end-users and application
	   designers.
	5. Can an end-user or application communicate to more than one database
	   at a time, in your system.


	D. Security, Recovery and Disaster management.
	   -------------------------------------------
	1. Describe how your system handles aborted transactions and how
	   read-consistency to the database is ensured.
	2. If your system makes use of snap-shot techniques, describe how
	   this is advantageously employed.
       *3. Describe the methods available in your system to backup and restore
	   a database. Is your system capable of on-line backup (while users
	   are using the system). Is INCREMENTAL backup supported.
       *4. Are tools available to ensure that all database files are backed up
	   together. Please describe your method.
       *5. Describe the tools available for the bulk loading of data. Do these
	   tools understand constraints defined on the database in order to
	   ensure no incorrect data gets loaded.
	6. What journalling facilities are available in your system.
	7. What level of security is provided to prevent unauthorised access
	   to your system and subsystems.


	E. Clusters and Networking.
	   ------------------------
       *1. Describe the operation of your system in a cluster or network
	   environment, especially how system recovery is managed in case of
	   node failure.
       *2. Does your system supports FULL Journalling in a cluster
	   environment.
	3. Is your system capable of distributed data management. Please 
	   elaborate. Is it possible to read/write data at remote sites.
	4. How many VAX Cluster nodes can your system cope with and can it 
	   utilise VAX Symmetric Multi Processing.
       *5. Does your system support two_phase commit in a cluster environment.


	F. Tuning and Performance.
	   -----------------------
       *1. Describe the facilities your system has available for fine tuning
	   the database.
	2. Describe the facilities available for on-line monitoring of system
	   performance.
	3. How many users can your system support.
	4. Please provide some statistics or other information on how your
	   system perform in the context of large databases, ie of 1 gigabyte
	   and upwards. What is the size of your largest database to date.
	   What is the maximum database size your system can support and
	   can this size be backed up within 8 hours (one working day).










				Page 4

	G. Tools for end-users, application designers and database
	   administrators.
	   -------------------------------------------------------
       *1. Describe the tools available in your system for each of these
	   catagories.
	2. Describe the tools available for fast prototyping.
	3. Describe the tools available for fast add-hoc queries to the 
	   database.
	4. To what extent does your system support the SQL ANSI standard.
       *5. To what extent can your system interface to the tools provided by
	   third parties.
	6. To which third generation languages can your system interface.
	   Which pre-compilers do you provide.
       *7. Does your system support dynamic SQL (SQL commands generated at
	   run time).
	8. Does your system interface to CASE tools. Please elaborate.


	H. Data Dictionary.
	   ----------------
	1. Can your system interface to a data dictionary or does it have a
	   intergral data dictionary.
       *2. Describe the features of your data dictionary:
	   Is it active or passive.
	   Is it capable of usage tracking and impact analysis.
	   Is it a distributed dictionary.
	   Is it capable of field level data definition.
	   Has it automtic relationship support.
	   Has it got a call interface.
	3. Describe its data security and integrity features.
	4. Describe the data model on which your data dictionary is based.


	I. User documentation.
	   -------------------
       *1. Describe your approach to user documentation. To what extent do your
	   user manuals enable end users and application designers to quickly
	   and easily learn how to use your system.
	2. What documentationn do you have available for database 
	   administrators and database tuners.
       *3. What in-depth technical documentation do you have available on the
	   internal operation of your system on order to assist in efficient
	   application design.

    
    Mike, I've got the new Database Features/Functions Grid but its
    not very different from the old one. (Heck, it seems to be exactly
    the same thing!). Cynthia should update this (ie. remove the version
    numbers from vendors that are shipping that exact version and just
    replace the matrix with a Y or N -Oracle for all intent and purpose
    is at V6.0).
    
    Something that I also found very useful was this PaineWebber report
    on Oracle corporation from April 1989 (Ken McMullen gave it to me).
    This is really useful. It is an independent study/audit of the ORacle
    corporation and the main thesis is around the parallels between Oracle
    in 1989 and Cullinet in 1985 (When IBM started bundling DB2 with
    the OS). These paint a very bleak picture for Oracle, and predict
    that DEC's bundling of Rdb may freeze the market like IBM did with
    DB2.

    As soon as Oracle quits selling V5 we will do that. Currently, Oracle
    usually pushes V6, but in the case of VAXclusters, they still attempt
    to sell V5. The bulk of the installed Oracle base has failed to upgrade
    from V5 to V6. Thus the need to show features of both versions.
    
    ---- Michael Booth

    Re: .3 Paine Webber
    
    Bhagwant, be very carefull about quoting from this report. It is
    old in financial analysis terms and ORACLE has released actual FY
    numbers since it was produced.

    Bhagwant, sorry I steered you wrong in the location of the Data
    Management Selection Guide. I talked to the Project Manager, Barbara
    Karten (DTN 297-6051) who pointed me too a NOTES conference with
    the locatons. The location is CSENG::BEPG$PUBLIC:.
    
    Hope this helps
    
    Paul
    


    
           <<< CSENG::$1$DUA3:[NOTES$LIBRARY]DM_FIELD_GUIDE.NOTE;1 >>>
                        -< Data Management Field Guide >-
================================================================================
Note 3.0               BEPG$PUBLIC - DM Field Guide Files              2 replies
CSENG::JANSEN                                        28 lines   8-DEC-1989 14:01
--------------------------------------------------------------------------------
    
    All documents pertaining to the Data Management Field Guide will be
    kept in the public directory: BEPG$PUBLIC:.  This topic is reserved for
    listing the files as they become available.  
    
    Files that end in .ps are postscript files and need to be printed on a
    postscript printer (LPS40, LNO3R).  To print a file on a postscript
    printer type:
    
    Print BEPG$PUBLIC:Filename param/data_type=postscript/queue=queuename
    
    The files with extension .TXT are text files.  These files contain only
    the text of the complete postscript file.
    
    Here is the list of files currently in the BEPG$DIRECTORY.
    
    DMFG_PROPOSAL.PS      Data Management Field Guide Proposal (Updated)
    DMFG_PROPOSAL.TXT    
    
    SALES_UPDATE.PS       Mock Up of Data Management Field Guide	
    
    Field_Guide_Proposal.PS  Original Data Management Field Guide Proposal
    Field_Guide_Proposal.TXT 
    
    Field_Guide_Pre_Project_Plan.ps  Preliminary project plan
    
    DMFG_Questionnaire.PS  Data Management Field Guide Questionnaire
    
    Enterprise_Data_Warehousing.PS  Data Warehousing Report
    
    
    

    re 546.2
    
    Good questionnaire...would it be possible to post your responses
    to the questionnarire?
    
    How about asking your customer for the responses of the other vendors
    and posting them also!
    
    The sales situation that Bhagwant is refering to is very interesting.
    It was decided before the RFP that Oracle was going to be the
    tools/database vendor. We have gone back in to talk to the customer
    about replacing Oracle with Digital tools/database. I took Oracle's
    RFP response (at the customer's request) and worked out Oracle's
    5 year costs (using the first year's maintenance for all five). Oracles's
    cost for 5 years was 650 K (with lots of discounts and they quoted
    there maintenance on discount price, yet the verbage said it is
    calculated off of list??).
    Our prices were 138 K (list - don't tell the customer) and providing
    more features and functionality than the products the big O quoted. By 
    the way the hardware platform is a cluster consisting of 2x6310 and 1x6210.
    
    We may or may not get the database business, but Oracle ain't going
    to get 650 K either.......it's too bad about the Oracle sales rep.
    He is a nice guy who used to work at DEC.
          
    signed:
    
    waiting ever so patiently for those "real" PC database tools...,
    
    Ken M.
    
    Ken McMullen

    re: -*
    
    Thanks for the input. Johannes, it would be nice to get the answers
    to those questions. If you could mail them (paper mail) or post
    them here it will be greatly appreciated.
    
    Also, thanks Paul. Actually, I am only using the PaineWebber report
    for your basic FUD about Oracle, and its working great. The customer
    knows the IBM/DB2 situation and in fact sees the truth about Mike
    Booth's assertion about market share PLUS penetration when it comes
    to 3rd party applications etc.

    Incidentally, we have educated the customer to the point that at
    least in one incident, the top level manager walked out of an Oracle
    schpiel and does not return their calls! Yow-sa!!