Conference ulysse::rdb_vms_competition

A few ideas...

�VAX is Oracle. There seems to be unrest in the Oracle camp and would probably
�be open to changing providing we could make a good business case.

So the customer is currently unhappy with ORACLE? In what areas? If you
concentrate on these then your chances are better.

�Their corporate systems are IBM with IMS and DB2. I don't see any move in
�changing that position.

Why doesn't the customer use Oracle in their corporate systems? If their
approach is open systems then this fact contradicts their approach. If the
customer wants the hardware vendor to also supply their business-critical
software then I see no difference on the VAX platform.

� Why should I devote the time, energy and money (sales support resouces)
� to move the customer to Rdb?

Because if you don't, then sooner or later your customer will need to upgrade
their VAXes. At that time ORACLE may suggest that they move to another platform
and you'll be looking for a new customer!

�Please don't view this note as heresy. I have really been struggling with the 
�answer.

It's not heresy, it just indicates that you're thinking about the issues
involved rather than blindly following Corporate strategies. 

John.

    Good points made by John. Sooner or later your customer will have to
    decide between DEC and Oracle. It's just a result of our strategy and
    theirs, we simply get further apart day by day. I've seen situations in
    Holland where Oracle aggressively tries to replace the VAX by other
    hardware.
    
    As for the revenue for Digital point of view, if your customer stay's
    with Oracle you will hardly sell any
    - dictionary tools
    - case tools
    - services and training
    
    
    Go for Rdb!
    
    Henny Olthof, TP-DB Holland

    Give your customer a Database Proprietary Information Disclosure.
    Digital recently announced (Nov 5th) it's long term database strategy
    called the "Information Network".   The Information network announcement 
    described Digital's intent to deliver a distributed database that runs
    across multiple platforms, and integrates hetrogenous data managers
    from other vendors.  
    
    In a time when so many database vendors are actively selling futures,
    I think that it is critical that your customer understand that
    Digital's database strategy is not VMS only.
    
      

	I've delievered a couple of the PIDs and the customers liked them

	How about another angle:

	Digital has much more to offer than database software and a few
	flashy tools.  When a customer buy's into Oracle's strategy,
	they only get pieces of the pie.

	When a customer buy's into Digital's strategy, they get ALL the products
        and services they need to be successful.   There are many pieces to
	the pie that all must work together for a quality solution.  Oracle's
	lack of support for VAXclusters, our TP monitors, and our dictionary
	strategy illustrates the just a few of the obvious missing pieces.

	But the real issue is this:

	If the account team does not list Rdb replacing Oracle on their account
	plan, your desire to do what's right for the customer may be wasted
	effort.

	-- gerry

	

>>Oracle's lack of support for VAXclusters, our TP monitors, and our dictionary
>>strategy illustrates the just a few of the obvious missing pieces.
    
    This is, of course, a reason why WE shouldn't recommend ORACLE, as 
    opposed to why a customer shouldn't but it :-).
    
    The main reason for not having ORACLE in your account is that they have
    a definite plan to unseat VAX/VMS from accounts and replace them with
    Sequent.  Their sales reps get bonuses for this.
    This information was confirmed by an ex-Oracle sales rep who recently
    joined us.
    
    Of course the company will sweetly say that they just let the customer
    decide which platform to choose.
    
    
    But back to the question of why Rdb (as opposed to why not Oracle):
    	Try the PID
    
    Cheers, Martin

    Let's be carefull withe the Sequent-bonus statement on Oracle sales
    reps. An ex-Digital, ex-Oracle and current-Digital salesman assured me,
    that in Holland there's no such thing as a bonus from Sequent if they
    (help) sell equipment.
    
    Must be at least country-specific then.
    Henny Olthof, TP-DB Netherlands

    re .4
    
    >	the pie that all must work together for a quality solution.  Oracle's
    >	lack of support for VAXclusters, our TP monitors, and our dictionary
    >	strategy illustrates the just a few of the obvious missing pieces.

    
    A Gartner Group Consultant I mwet with last week said that the new
    Oracle version DOES work properly in clusters. I hope you can verify
    this.
    
    Barry
    

    Barry,
    
    The Gartner Group consultant either does not understand "VAXcluster
    support" or has seen some unreleased/un-announced version of Oracle.
    
    Don't forget Oracle usually re-defines the term when they do not meet
    the industry definition. Look at what Oracle calls a trigger and
    referential integrity - a piece of code attached to a form!
    
    Ken McMullen

    Sure, my customer say's he saw Oracle demonstrated on clusters (release
    expected soon). BUT WHAT DOES THAT MEAN:
    - auto-failover
    - database recovery
    OR DOES IT MEAN THAT YOU CAN INSTALL ORACLE ON TWO OR MORE NODES IN A
    CLUSTER AND IT DOES NOT CRASH? (like in version 5)
    
    Henny Olthof, TP-DB Netherlands

    
    ORACLE isn't the only RDBMS out there.  Remember INGRES (that's
    our ULTRIX/SQL)?  Thing is a fairly elegant multi-server, multi-
    threaded architecture and is ported to 3 of DEC's operating systems
    (RISC/ULTRIX, VAX/ULTRIX, and of course VMS).  The full db engine
    INGRES 6.2 is bundled with ULTRIX; INGRES current release, 6.3,
    should be out with ULTRIX 4.1.  
    
    So, for customers with a heterogeneous environment, seeking a single
    product that runs across all DEC platforms, our group (Govt DCC,
    Landover, MD) finds INGRES-ULTRIX/SQL a nicely priced solution
    for some customers.  INGRES tools and applications sw are included
    in DSPS, too, so we get certs credit for selling those. 
    
    Elaine
     (who wishes so many Govt. RFPs didn't specify UNIX)

    
    Oracle have announced their V6.2 product which according to the
    announcement supports VAXclusters.  The announcement dated late
    November said that they expect the product to enter Beta test at
    selected sites in December.  Have a look at the last paragraph, it
    gives new definitions to marketspeak.
    
    
    Some statements from the article:
    
    "Oracle 6.2 Development Manager Andy Laursen said the VAXcluster's lack
    of shared memory was an obstacle in providing the 'fast commit'
    features introduced in V6.  Laursen said this obstacle caused too many
    lock operations which was the reason for the denigration of
    performance.  V6 was tested for speed of lock operations and required
    200 microseconds for each operation."
    
    "However, it was found that as other CPUs were added, V6.0 performance
    was not scaling as expected and the Distributed Lock Manager problem
    was discovered.  Oracle solved the problem by implementing a
    distributed cache accross the loosely coupled VAX nodes.  The new
    implementation improves performance by decreasing the amount the
    Distributed Lock Manager calls."
    
    "Laursen said the new version on VAXclusters would be particularly
    advantageous for users in occasions with query only and partitioned
    data."
    
    "While the new version is seen as a fix to problems not anticipated by
    Oracle, Jernigan said 'VAXclusters create problems for all
    relational databases'.  'Most can't do parallel cache management, which
    means they can't have the buffer cache distributed in several pieces
    accross several nodes of the VAXcluster,' Jernigan said.  'They run
    into the same Locking Manager problems and do different thimgs to get
    around them.  Some don't have a cache at all, so they do lots of I/O
    and are really slow.  Others have a cache but don't use it well, do
    lots of lock operations, and again are slow.  We managed to minimize
    I/O and lock operations.  It was more of a re-architecture than a fix."