Subject: | |
From: | |
Reply To: | |
Date: | Fri, 4 Apr 1997 05:58:07 -0800 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
John Dunlop wrote:
>[This note has been sent to the following InterNet address(es):
>[log in to unmask]]
>
>
>Jeff Kell wrote :
>
>[snip]
>
>>The bottom line here is "authentication".
>
>>Let's suppose you are on a 3000 (nothing else for now). You may have:
>
>[snip list of passwords etc]
>
>>That is a *lot* of redundancy/replication, and it isn't an inclusive
>>list by any means. And these are *ALL* on *ONE* system, let alone being
>>an enterprise solution.
>
>>DCE promises some distributed authentication (with programmatic access
>>by applications) but hasn't exactly taken off like a rocket.
>
>I have been researching the use of DCE on the HP3000 for use as an
>authentication and authorization method for securing logons to multiple
>systems. DCE/3000 uses the Kerberos 5 authentication model which requires a
>single username/password logon which is validated by a security server and
>returns a token to the user. This token is used to access other systems in the
>DCE "cell". It appears that DCE/3000 (which is based on OSF DCE 1.0.2) is
>dropping behind the versions on other platforms (e.g. HP-UX, VAX etc) and HP
>have not confirmed that there are any plans to offer an upgrade. It is driven
>by customer demand and apparently not enough customers are driving it. I am
>hoping they will port it from HP-UX but it needs other people to request
this ac
>tivity.
Ah - the chicken and the egg. As the HP3k becomes more mature and the funding
models are setup to see a direct return on each investment, technology
like DCE and CORBA will not become mainstream MPE products.
Why not? Well, HP can justify not keeping up by saying "We don't see enough
customers asking for these products so we won't port or we won't extend what
we already have ported".
This is failed thinking for the following reasons:
1. The typical MPE installed base customer probably doesn't even know
what these technologies are and can't see why they would want them.
HP needs to take a leadership position with regard to certain technology.
2. It keeps the HP3k from playing big-time in the distributed computing
arena. Too bad. HP has one of the best CORBA ORB's (Object Request Broker)
running on HP-UX. They have recently announced support for linking
DCOM (Microsft's Distributed object technology).
3. It continues to provide an entry barrier for more advanced technology and
software products being ported to the HP3k.
This is where I should chime in and begin discussing additional ideas for
adjusting the MPE funding model [look for a posting soon on this topic] ;-)
Its not that I don't have something to offer, I just have to leave any
minute to go install some software on a *new HP 3000*.
Yes that's right - a new customer who is joining the fold of HP 3000/MPE.
And why? Because they bought sofware from a vendor (us) that *only runs
on the HP 3000*.
And that is because the HP3k offers the best value proposition for our
target customer base.
But, it would offer a much higher value proposition if I could develop
distributed objects for the HP3k that could be instantiated from my
Windows'95 Visual Basic or Java desktop applications. if only. sigh.
duane percox
|
|
|