Enron Mail

From:bob.shults@enron.com
To:travis.mccullough@enron.com, andrew.edison@enron.com
Subject:HoustonStreet Testing
Cc:andy.zipper@enron.com
Bcc:andy.zipper@enron.com
Date:Thu, 17 May 2001 09:14:36 -0700 (PDT)

Pursuant to the HoustonStreet Price Posting Agreement and subsequent amendment, the companies have conducted extensive Interface testing. The testing has been an iterative process over several weeks and has included both stress testing and unit testing. The testing was conducted to ensure interface functionality and load performance.

On May 7 - 8, 2001 we tested with HoustonStreet to determine whether the Sponsor Interface Acceptance Criteria had been met. Based upon the following results I do not believe that all the criteria have been met and recommend against putting the Interfaces into production.

1. Although the Sponsor was able to accept continuous price updates we experienced a 1 - 8 second time variance between price updates on EnronOnline and price updates on HoustonStreet. Testing periods of up to 1.5 hours were included in the testing. Price updates during the testing ranged from 20 - 40 updates per second with peaks of approximately 60 updates per second. The 1 - 8 second delay on the HoustonStreet website will result in a high percentage of failed transactions with Enron.

2. No testing was done at 100 -200 updates per second.

3. HoustonStreet was unable to do a three day uninterrupted period to allow for testing. As a result of incidents seen in the testing on May 8, 2001, HoustonStreet was still making adjustments to software as evidenced by the May 11, 2001 email from John Whyman of HoustonStreet. Without a three day "quiet" period we are unable to finish testing the functionality of the interfaces. Until such time as the software and hardware are locked down, we cannot be assured that changes that are being made are not impacting the results of previous testing. Essentially without a lock down period we could not be sure that any functionality would work when we went live.

4. HoustonStreet did provide an excel matrix for displaying the Enron product descriptions. Enron requested that the matrix be prominently displayed on the Sponsor Platform. We were unable to view the matrix on the Sponsor Platform. If the matrix was not prominently displayed on the HoustonStreet website, customers transacting on HoustonStreet may actually buy or sell a product that is different than they thought. The matrix was required to ensure that anyone transacting with Enron would know exactly what product they were dealing with based upon the EnronOnline short description.

5. Functionality testing could not be completed due to the inability to do a three-day uninterrupted test as required in number 3 above. As long as changes to the software and hardware continued Enron can not be assured that changes would not impact system functionality. We had planned to do a test of all functionality during the three-day quiet period. Testing of system functionality that was conducted on May 7 - 8 revealed the following critical issues:

We saw incidents of incorrect Enron price updates posting on the Sponsor Platform. One specific examples included a crossed market display with a bid price of $6.00+ and an offer price <$1.00
We experienced instances of both successful and failed transaction messaging that were delayed anywhere from 10 seconds to several minutes.
We experienced successful transactions with Enron with no corresponding responses back to the Sponsor Platform.
We experienced instances where price updates were causing products to appear and disappear. HoustonStreet indicated that the issue was the result of incorrect mapping.

Given the functionality problems we witnessed, a disconnect between Enron and the customer as to what transactions have been entered into would occur. This disconnect would result in substantial price exposure during the period that the trade is not confirmed. In addition, given the timing of the confirmations (positive and negative) a customer might be confused as to which confirmation was associated with which transaction attempt. This situation once again would result in exposure from price movements in the markets.