ne of the largest changes in SQL Server 2008 Reporting Services (SSRS) is that it no longer relies on IIS. Due mostly to feedback from end users on security (most do not like to keep web hosts active on their database servers), the added benefits include reduced memory requirements and service spool-up time.
Testing the Performance of Requests
The tests used for this article were designed to time the performance of requests against the Reporting Service web service (ReportServer2005.asmx). Report rendering and data retrieval was not tested.
The performance tests were conducted using virtual servers running both SQL 2005 and SQL 2008 (CTP4). The test harness was built in VS 2008 and run from the host machine, which was then connected to each Virtual Hard Disk (VHD) (via a web service) when needed. See the sidebar for the specifications and applications.
Both virtual machine?s (VM?s) were running in parallel for the duration of each test. Web service functionality was evaluated using three separate tests:
- Cold Server Spool Up
- Warm Refresh
- Warm Server Spool Up
A simple ASPX test harness running on the host laptop OS made three connections to ReportServer2005.asmx:
- Accessed the web service on the tested VHD via URL
- Opened a connection to read a report definition from the web service
- Looped through the list of parameters in that report
|Author’s Note: SQL Server Katmai CTP4 currently uses ReportServer2005.asmx; a ?2008? version is not currently available, though it works against Katmai DBs.
In Visual Studio Web Developer Express 2008 (a free download from Microsoft), the author created a new project and a proxy to the Reporting Service 2005 web service, and named it RS2005. Listing 1 shows the code for the default.aspx.
Each test was provided three trials. If an anomaly occurred, several more tests were run to minimize the effects of running all the tests on a laptop. At one point, the SQL 2005 VHD timed out after 90 seconds on a cold boot test, but after several successful trials it did not seem fair to keep that test in. The tests were repeated until the results were consistent; you could eventually predict the length of time with fair accuracy for each test, as there was not much discrepancy.
Test 1: Cold Server Spool Up
In the Cold Server Spool Up test, each VHD was restarted and booted into an administrative screen. The ASP test harness accessed the VHD, simulating a first web service request to a freshly booted Reporting Services (RS) Server.
Perhaps due in part to the omission of IIS reliance in Katmai, SQL Server 2008 wins this test hands down. The average web service test in SQL Server 2008 took a little over 36 seconds, while web service access to a SQL Server 2005 machine took about 66 seconds, or an improvement of about 45 percent (see Table 1 and Figure 1)!
Each Warm Refresh test made an additional call to the RS Web Service within a few moments of the initial test call to simulate another user accessing a report prior to a reboot or any service/application pool refresh.
After several attempts the SQL 2005 results were fairly predictable at under 0.3 seconds; however, the Katmai machine exhibited a strange anomaly: response time was either well under 0.1 seconds (twice the speed of 2005) or as high as 2.9 seconds. This was reproduced several times: two trials at 0.15 seconds, the next at 2.8, another trial at 0.1, and the next two at 2.9 and 3.0. This could have been simply the result of the VHD test environment (antivirus, etc.), but because this behavior was not observed with SQL 2005, one of the high values from the Katmai test were left in the test results (see Table 2 and Figure 2).
The Warm Server Spool Up test was meant to simulate a user accessing the web service after it had been idle on the server for some period of time. The RS?s reliance on IIS means it is subject to that system’s control of system cache and application pool resource management. After periods of inactivity it is common for a host to wind down. Still, a user request on this host may not take as long as a cold spool up. To simulate this, the application pool controlling the web service in IIS Manager was recycled.
There was little reason to test the SQL 2008 machine as IIS was absent, so presumably that same refresh should never occur. As there is no IIS control, the author predicted that this service would never actually stop except due to error or server reboot. Just for fun, the author wound down the Windows service controlling the SQL Server 2008 Report Server anyway. Even with this imposed handicap, Katmai was still a little over 1 second faster, an improvement of about 11.5 percent. The real story here is the nearly complete elimination of the time lag due to application pool reset (see Table 3 and Figure 3).
Overall, the clear winner here is Katmai, SQL Server 2008. If performance, scalability, or IIS security held you back from SQL Server in the past, now is the time to begin testing with Katmai SSRS.
Comparing SSRS Development Efficiency in SQL Server 2008 Katmai to the SSRS 2005
|Figure 4. Report Designer: Here is the Report Designer Preview?with ribbon!
Unfortunately, all is not peachy in the land of Katmai. The next test included development in CTP4 using the Report Designer Preview. While still clearly a prerelease product (there is no expression builder yet so you have to write code in a simple text box), it hints at the new direction that the RS team is taking. This new version takes bits from BIDS, Office 2007 (ribbon included), and Report Builder 2005, and mashes them together for a single development interface (see Figure 4). The new interface should help capture hard-core report developers who felt Report Builder was too simplistic.
Report Builder will not be going away (at least not yet) but it is clear that Microsoft intends business analysts and non-techies to use the ?Report Designer.? and this is where Katmai falls short.
New forms and pop-ups take an Office 2007 and Vista approach; they are clean and functional. But the functions that work for single clicks (like bolding in a Word document), do not easily translate for development with much greater form interaction.
A case in point is the sheer number of additional clicks it takes to build a report in RS 2008. Creating parameters used to take place on a single form, now a popup modal form appears for each new parameter. That popup has very nice visual appeal, but the design loses tons of efficiency. All properties used to be set on a single interface (parameter name, defaults, valid list of values, etc.).
To call the previous form design of RS 2005 ?Spartan.? would be an overstatement, but it got the job done. Conversely, to set values on RS 2008 Report Designer you must click into each section: General, Available Value, Default Values, and the new Advanced. There is no shortcut key, so you will be jockeying from keyboard to mouse for these extra steps.
|Figure 5: RS 2005 Parameters: This figure shows a simple grid that easily allows entering values.
As previously mentioned, the Katmai version is clean. But after you get past the color gradations a few challenges arise. For example, the list of parameters display only on the General tab. If you are working on the Available Values tab you can no longer see your parameters. Trivial? Perhaps. But many people like that kind of visibility.
Further adding insult to injury, it may take up to five times longer to provide valid lists of “Available Values” than it did in the earlier version of Reporting Services. In RS 2005 always presented you with a simple grid in which you could easily enter values (see Figure 5). And the grid dynamically resized; for each new value, a row was automatically created.
|Figure 6: The Grid is Gone in 2008: This figure shows adding Report Parameters in SQL Server 2008 (Katmai).
In SSRS 2008, as Figure 6 shows, the grid is gone. Now you must add each possible value by clicking an Add button, and then clicking into the row to provide the detailed elements. A significant level of efficiency is lost. Katmai uses this long-winded approach for every input, from parameters to data sets and properties. Get used to it.
Dropping the automatic grid function on development UI components in favor of the new Add button concept takes a huge step backwards in productivity. Of course, you are not required to manually enter values if you derive them from a query. But sometimes values are entered during the prototype phase or when it is unnecessary to roundtrip to a database.Please keep in mind that many items may change before the retail release. However, it does seem clear that the behavior described in this article is expected in the upcoming version.
See sidebar 2 for the Author?s Parameter entry wish list for future versions.
SSRS Reaches Adolescence
SSRS is growing up, that?s clear. With this third version the product keeps maturing. Benefits will be realized in performance and scalability and similar improvements are expected in SSIS. However, the new development UI does little to embrace the developers who have been using this tool since the 2000 version. No doubt, new converts to SSRS will adjust nicely to the Office 2007 look-and-feel, but business users and other non-techies might feel as though they have been left hanging.