Does SQL Server 2000 support dynamic clustering like
Oracle 9i?

What does this mean: Microsoft's documentation, "SQL
Server 2000 does not support this type of clustering"?

Does SQL Server 2000 has the same expansion capabilities
as Oracle 9i?

Is SQL Server 2000 more Susceptible to Virus attacks than
Oracle 9i?

Cassandra Kenyon
Microsoft Certified Attain Consultant
Aston Business Solutions
7050 Jomar Drive
Whitmore Lake, MI 48189-8241
Direct Phone: 734.997.2058

Answers Inline

"Cassandra Kenyon" <Kathy.Nohr (AT) astongroup-us (DOT) com> wrote

No. SQL Clustering is a failover ONLY solution, not a scale out solution.
To overcondense, Oracle clustering is predicated on a particular transaction
isolation method that is not part of SQL 2000. Dynamic clustering is not
linear in performance either. The overhead of managing a distributed lock
system eats a lot of system bandwidth If you want more power with SQL, buy
a bigger box. SQL is designed to scale UP, not out.


TCP-C benchmark back and forth every few weeks for a while now. Of course,
the actual numbers are way above anything any normal database will ever use.
On similar hardware, the numbers come out pretty close. It is hard to slip
anything past a few hundred developers regardless of whether they work in
Redmond or San Jose.

Depends. Windows based systems SEEM more vulnerable because any fool can
set one up. Most actually try. If the system (network, host, app, etc) is
properly designed, installed, and maintained, they are about even.

--
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
CareerBuilder.com

Oracle 9i RAC or Oracle 10g implementes a shared disk
cluster model. In theory, the implementation is very
appealing in that you have both high availability (i.e.
failover) and dynamic load balancing. It's debatable (or
not depending on who you talk to) to what extent the model
can scale. The Oracle clustering implementation is similar
to DB2 shared-data clustering on zOS, though the latter
has a different way of distributing cache and managing its
coherency among the nodes. I don't believe DB2 UDB -- the
baby DB2 -- implements shared-data clustering.

The model of SQL clustering is shared-nothing. No
computing resource is used by more than one node at any
given time, other than the fact that the nodes/ are
interconnected. But I wouldn't say one is
necessarily 'truer' clustering. They present you different
choices in satisfying your requirements. With Intel
servers getting very powerful, your applications may not
need all that dynamic load balancing -- again a debatable
statement.

It's safe to stay away from arguing for one or the other
because doing so often stirs up a lot of emotion :-) But
you can find a lot of materials comparing/contrasting the
two approaches by googling the web. In many cases, the
differences between these two clustering approaches
probably wouldn't matter that much because the other
factors often turn out to be more important in whatever
decisions you are trying to make.

Linchi