Microsoft Time Zone Data Update Tool and daylight-saving time changes in 2007
The Microsoft Time Zone Data Update Tool (which was also referred to as Time Zone Update Tool), is a program provided by Microsoft to automate the process of “rebasing” events in Outlook and Exchange calendars. RIT conducted internal testing on several scenarios involving this tool before deciding whether to apply it globally to the calendar data in all RIT Exchange accounts.
Methodology
ITS ran tests to determine the effects of the Microsoft Time Zone Data Update tool. The goal was to determine whether there is a benefit or drawback to running the Time Zone Data Update Tool across all of the Exchange mailboxes hosted by the university. Again, this tool was only meant to automate the proces of “rebasing” events solely within the extended daylight-saving time period:
- between March 11, 2007 and April 1, 2007
- between October 28, 2007 and November 4, 2007.
Microsoft’s guidance was to install the Windows operating system and Entourage client patches, and run the Time Zone Data Update Tool immediately — conducting all of these steps in quick succession. We needed to simulate situations where these steps had not been run in quick succession, since the patches were widely available for installation by RIT faculty, staff, and students well before we felt we could run the the rebasing tool enterprise-wide.
While we used the Outlook version of the rebasing tool in the study, the results should exactly follow those from the Exchange edition, which acts a wrapper around the Outlook tool.
We primarily studied two test matrices. The first was a four-cell matrix of Microsoft Windows clients, running the same version of Outlook 2003, with two axes:
- before and after the Windows operating system patch
- before and after the Time Zone Data Update Tool had been run (from a patched Windows client).
The second matrix was an eight-cell one covering Apple Macintosh clients:
- Mac OS X 10.4.4 (before the time zone rule update) versus Mac OS X 10.4.8 plus the February 2007 DST Update, each further subdivided by the use of Microsoft Entourage 2004 11.3.2 versus the patched 11.3.3
- before and after the Time Zone Data Update Tool had been run (from a patched Windows client).
Various categories of test events were created in fresh test accounts on RIT’s Exchange 2003 servers, which had previously been patched for the new DST rules. For Windows and Outlook, the events were created by hand. In the case of Entourage 2004, the test events were created via AppleScript. Additional scenarios involving Outlook Web Access were also tested, and events were created manually in both Microsoft Internet Explorer 6 (OWA Premium) and Mozilla Firefox 1.5 (OWA Basic). The events were:
- Standalone one-hour events before, during, and after extended daylight-saving time
- Standalone 12-hour events that overlapped into or out of extended daylight-saving time
- All-day events before, after, and during extended daylight-saving time
- A recurring one-hour weekly event that began before extended daylight-saving time, continued through it, and ended after it.
We could not test every scenario, primarily due to time constraints and the size of the test matrices. We didn’t specifically study event invitations, resource booking scenarios, or Public Folders. We also didn’t explicitly study events viewed between clients on various platforms, so we did not cover situations like, “An Entourage client views a calendar predominantly edited in Outlook.”
Results
After extensive testing of the Microsoft-supplied Time Zone Update Tool, ITS has decided not to run it globally across the RIT Exchange environment. We have only found it effective in a limited set of cases which we do not believe reflects the campus today.
While individuals may choose to run the tool and Organizational Unit (OU) administrators may choose to deploy the Outlook version of it (especially if their environments match the limited set of cases where the tool is effective), ITS does not generally recommend using it to rebase Outlook calendars. Instead, we recommend patching the client operating system and applications, and manually correcting calendar events afterwards. This allows the most control.
This decision is due to the inordinate number of incorrect adjustments in meetings the utility made during our testing. Even under the most favorable circumstances — involving fully-patched Microsoft Windows XP systems, running Microsoft Outlook 2003 — 40% of the meetings moved by the utility were not moved to the correct time.
When tested with some unpatched systems the utility’s effectiveness degraded further. And, when tested with fully-patched Apple Macintosh systems using the latest Microsoft Entourage 2004 Exchange client, all of the meetings moved by the utility were moved in error.
Even though ITS will not run the tool, we must stress that patching your own computer is critical, particularly if you use the RIT Exchange calendar and plan to manually correct your own events.
Microsoft Windows and Microsoft Outlook
| Operating system and application | Before Time Zone Data Update | After Time Zone Data Update |
|---|---|---|
| Microsoft Windows XP SP 2 with Microsoft Outlook 2003 |
|
|
| Microsoft Windows XP SP2 plus DST patch with Microsoft Outlook 2003 |
|
Apple Macintosh and Microsoft Entourage
| Operating system and application | Before Time Zone Data Update | After Time Zone Data Update |
|---|---|---|
| Apple Mac OS X 10.4.4 with Microsoft Entourage 2004 v11.3.2 |
|
|
| Apple Mac OS X 10.4.8 plus DST patch with Microsoft Entourage 2004 v11.3.2 |
|
|
| Apple Mac OS X 10.4.4 with Microsoft Entourage 2004 v11.3.3 |
|
|
| Apple Mac OS X 10.4.8 plus DST patch with Microsoft Entourage 2004 v11.3.3 |
|
Summary
To summarize the test results:
- The Time Zone Data Update Tool did not move every event in a calendar — whether the operating system and/or application were patched or not. In our testing, it “rebased” few of our overall set of test events.
- Overall, in the various scenarios we tested with desktop Exchange clients, if the tool rebased an event, it was more likely to shift it to an incorrect time when viewed on a fully-patched operating system and application. There were successes, but we found them only in specific circumstances.
- The greatest benefit of the Microsoft Time Zone Update Tool for the RIT Exchange environment was to calendar events posted by unpatched Windows XP SP2 systems using Outlook 2003, and then viewed later by a patched version of Windows. This matches up with Microsoft’s recommendations for when to run and not run the Time Zone Data Update Tool. Based on Microsoft’s release of the Windows XP DST patch as a critical update in late 2006 and the extensive reliance on Windows Update and Microsoft Update on campus, it was likely that the DST patch was installed widely before we could run the Time Zone Data Update Tool, and thus this situation would not apply to RIT.
- The Time Zone Data Update Tool moved many calendar events to the wrong time when they were both posted and viewed by patched Windows XP SP2 systems with Outlook 2003. Again, this was in line with Microsoft’s recommendations about when to run and not run the Time Zone Data Update Tool.
- The Time Zone Data Update Tool was a detriment in every change it made to calendar events created by patched and unpatched Microsoft Entourage clients. Events that were moved were displayed afterwards at incorrect times.
- Based on these results, ITS has decided against running the Time Zone Data Update Tool globally on RIT Exchange accounts. Please see the RIT Microsoft Exchange and daylight-saving time changes in 2007 page for more details.
- The above listing of test results is only a summary. All test results have been posted on the wall outside the offices of Jeremy Reichman and Shawn Thomas in the ITS office area in the Gannett Building (7b). Anyone wishing to see these full results in poster form can stop at the ITS Service Desk during Institute business hours and ask for access to the daylight-saving time tests, which we plan to keep available until at least April 2, 2007.
 |
 |
