users@jaxb.java.net

Testing Pay.gov and the Verification Engine

From: David F. Wall <dwall_at_foliage.com>
Date: Thu, 22 May 2003 09:41:30 -0400

Since I'm the one with his hands on most of the threads, I'll be
coordinating. This message is soliciting:

1) Information on when the RAF machine in Redmond can be ready, and
information on connecting to it. (RAF)

2) Information on whether or not AppServer1 in Richmond can communicate with
an RAF machine in Redmond (Don, Roberta?)

3) Information on whether or not an NT-based Pay.gov instance in Foliage
would be suitable for conducting this test (John, Mark, Kevin, Sushama)

4) Information on how long it would take to get the NT machine outside the
Foliage firewall up and running with the 2.3 release candidate (Nathan,
Murray, Don)

Background:

We need to test a release candidate for Pay.gov 2.3 with a version of the
RAF Verification Engine hosted in Redmond. This is prompted by a call I
received at about 4:30 yesterday from RAF, and stems from a desire to find a
way to get this testing done before the new RAF deployment in Richmond is
complete.

Setting a Pay.gov instance up to use a Verification Engine hosted at a
different port and server is merely a question of changing the values of two
parameters and rebooting the Pay.gov instance. The more difficult question
to answer is whether or not a Pay.gov instance will actually be able to open
a socket to the configured host and port and conduct communications.

There are three possible configurations for conducting this test:

1) Reconfigure AppServer 1 in Richmond FT to point at the host and port for
the machine in Redmond and conduct the test there.

This is the ideal situation, but I am not sure the network configuration at
Richmond will allow AppServer1 to communicate with the machine in Redmond.
Answering that question requires knowledge I don't have.

2) Reconfigure a Solaris host at Foliage to point at the host and port for
the machine in Redmond and conduct the test there.

I mention this for the sake of completeness, but it really isn't feasible,
since it is not possible to move either of Foliage's Solaris hosts outside
the Foliage firewall to allow bi-directional communication with the RAF
machine in Redmond.

3) Reconfigure an NT-based host at Foliage to point at the host and port for
the machine in Redmond.

We already have a machine outside the Foliage firewall for such purposes,
although it would require software upgrades and an installation of the
Pay.gov 2.3 release candidate before becoming a suitable platform for
conducting this test, assuming that an NT-based host can be considered
suitable from a Quality Engineering standpoint.

If we can get at the machine in Redmond from AppServer1 in Richmond FT, then
as soon as RAF says the machine in Redmond is ready, we can get the host and
port from RAF, schedule time on AppServer1 to take the instance down, change
the parameters, conduct the tests, and then restore the instance's previous
state.

If that isn't possible, then we need to set things in motion to get a NT
machine at Foliage set up to serve as the Pay.gov instance.

Dave Wall
Foliage Software Systems