Comparing Enterprise Development...

Comparing Enterprise Development Productivity:

SQL Server 2005 versus

SQL Server 2000

3 Leaf, LLC October, 2005

Table of Contents

1.0 Executive Summary..............................................................................................................................3 2.0 Study Goals and Methodology ............................................................................................................4

2.1 Overview..............................................................................................................................................4 2.2 Development Firm: 3 Leaf ...................................................................................................................4 2.3 The Baseline Contoso Times Web Application ...................................................................................4 2.4 High-Level Overview of Development Phases in the Study................................................................6

3.0 Phase 1 Development: Paid Articles and Subscriptions ..................................................................7 3.1 Technology Requirements Underlying Phase 1 Development ...........................................................7 3.2 Summary of Phase 1 SQL Server 2005 Development........................................................................7 3.3 Summary of Phase 1 SQL Server 2000 Development........................................................................8 3.4 Phase 1 Competitive Outcome............................................................................................................8

4.0 Phase 2 Development: Targeted Advertising ....................................................................................9 4.1 Technology Requirements Underlying Phase 2 Development ...........................................................9 4.2 Summary of Phase 2 SQL Server 2005 Development........................................................................9 4.3 Summary of Phase 2 SQL Server 2000 Development......................................................................10 4.4 Phase 2 Competitive Outcome..........................................................................................................10

5.0 Phase 3 Development: Middle-Tier Caching ....................................................................................10 5.1 Technology Requirements Underlying Phase 3 Development .........................................................11 5.2 Summary of Phase 3 SQL Server 2005 Development......................................................................11 5.3 Summary of Phase 3 SQL Server 2000 Development......................................................................11 5.4 Phase 3 Competitive Outcome..........................................................................................................12

6.0 Phase 4 Development: Custom and Ad-Hoc Reporting..................................................................12 6.1 Technology Requirements Underlying Phase 4 Development .........................................................12 6.2 Summary of Phase 4 SQL Server 2005 Development......................................................................12 6.3 Summary of Phase 4 SQL Server 2000 Development......................................................................13 6.4 Phase 4 Competitive Outcome..........................................................................................................13

7.0 Phase 5 Development: Offline Smart Client Application for Ad Sales ..........................................13 7.1 Technology Requirements Underlying Phase 5 Development .........................................................14 7.2 Summary of Phase 5 SQL Server 2005 Development......................................................................14 7.3 Summary of Phase 5 SQL Server 2000 Development......................................................................15 7.4 Phase 5 Competitive Outcome..........................................................................................................15

8.0 Overall Outcome .................................................................................................................................15 9.0 Key SQL Server 2005 Usability Benefits Reported by Developers................................................16

9.1 Encryption Advantages .....................................................................................................................16 9.2 Reporting Advantages.......................................................................................................................16 9.3 Replication Advantages.....................................................................................................................16 9.4 Interface and File-Management Advantages ....................................................................................17 9.5 Common Language Runtime Integration Advantages ......................................................................17

10.0 Summary and Conclusion................................................................................................................17 Appendix: The Development Environments ..........................................................................................18

SQL Server 2005 Software Environment ................................................................................................18 SQL Server 2000 Software Environment ................................................................................................18

Comparing Enterprise Development Productivity: SQL Server 2005 versus SQL Server 2000

of 18

1.0 Executive Summary

Microsoft SQL Server provides features such as failover clustering, replication, database snapshots, database mirroring, and snapshot isolation, helping ensure that your critical databases maintain high availability. The platform has always provided outstanding programmability, enabling developers to build robust custom solutions efficiently, and those capabilities have improved with each successive generation of the product.

A key proof point about the relative programmability of SQL Server 2005 versus SQL Server 2000 is how efficiently each version allows developers to achieve the application functionality required to meet business needs. This productivity study compared two developers working independently to provide the same added features and functionality to a baseline application; one of the developers worked with SQL Server 2005, and the other used SQL Server 2000. By comparing the time required by each developer to perform equivalent development tasks, the study quantified one instance of improved development efficiency in SQL Server 2005, relative to SQL Server 2000.

Both developers in this productivity study worked in isolation from one another, from the same baseline application: a dynamic, database-driven web application that delivers news content, including registration for premium content, using an ASP.NET front end and a SQL Server back-end. The study was divided into five phases, with each phase adding a new feature set:

•

•

•

•

•

Phase 1, Paid Articles and Subscriptions, enabled the site to charge for content, including providing the ability for premium subscribers to syndicate content using a web service.

Phase 2, Targeted Advertising, provided premium advertising services, matching ads to content.

Phase 3, Middle-Tier Caching, increased the scalability of the site and decreased the database load with middle-tier caching.

Phase 4, Custom and Ad-Hoc Reporting, provided reporting functionality to gather metrics associated with the site's revenue-generating activities.

Phase 5, Offline Smart Client Application for Ad Sales, provided a tool to support sales while salespeople are online or offline.

The study was designed to keep the applications in alignment in terms of architecture, UI elements, and database structure throughout these five phases, in order to make the productivity study as fair a comparison as possible between the platforms under study. Communication between the developers was intentionally severely limited and moderated by a project manager to further isolate the effects of either development effort on the other.

The results of the study were that the SQL Server 2005 developer benefited dramatically from the built-in functionality and automation afforded by SQL Server 2005 that was not available to the SQL Server 2000 developer. Features such as the built-in encryption, reporting, automated replication, and various levels of additional functionality in SQL Server 2005 such as those related to interface and file management, as well as Common Language Runtime Integration created efficiencies for the SQL Server 2005 developer. Those efficiencies caused the SQL Server 2005 development in the study to be completed far more quickly than the development effort based on SQL Server 2000, as shown in the table below:


of 18

Phase SQL Server 2005

SQL Server 2000

Key SQL Server 2005 Feature

Phase 1: Paid Articles and Subscriptions 27.5 hours 98 hours Built-In EncryptionPhase 2: Targeted Advertising 32 hours 85 hours Built-In EncryptionPhase 3: Middle-Tier Caching 6 hours 6.5 hours Interface AdvantagesPhase 4: Custom and Ad-Hoc Reporting 16.5 hours 29.5 hours Report BuilderPhase 5: Offline Smart Client Application 15 hours 50 hours Automated ReplicationTotal Development Time: 97 hours 270 hours

The SQL Server 2005 developer was able to complete the tasks called for in all five phases of the study in a total development time of 97 hours, or just over 12 developer-days. The SQL Server 2000 developer reported total development time of 270 hours, or just under 34 developer-days, approximately 2.8 times the amount of time required by the SQL Server 2005 developer.

2.0 Study Goals and Methodology

2.1 Overview The goal of the study was to provide an objective measure of developer resources required to obtain common functionality using SQL Server 2005 versus SQL Server 2000. Developers were chosen for each platform with similar backgrounds in SQL Server and web-application development, and each of them worked in isolation from the other to add the same specific functionality to a common baseline application. Further detail about the developers in the study is contained in the appendix to this document.

In order to provide a means of comparison between the parallel development efforts, the study was divided into five phases, as described below. Each developer kept a running log of the hours they invested in the development of their solution, as well as development diaries to be used in the preparation of this study. Those logs and diaries were not visible to the other developer during the course of the project. Further, in order to maintain a level playing field and independence between the developers, only limited communication was allowed between them, as moderated by a project manager. Those discussions were limited to answering questions that were applicable to development in both environments. Developers were free, however, to conduct research using other resources such as web sites, newsgroups, etc., just as they would for a real-world project.

2.2 Development Firm: 3 Leaf The development firm that carried out this productivity study, 3 Leaf, was chosen for its extensive experience working with both mature and emerging software technologies and products from Microsoft and others. Since 2000, 3 Leaf has specialized in creating reference, proof of concept, and sample applications, targeted training content, and demonstrations and keynote presentations for Microsoft, Intel and others, in addition to more traditional development efforts. Its developers have been early adopters of each generation of SQL Server, ASP, and ASP.NET, helping the new aspects of these technologies to become adopted into the mainstream. This productivity study continues with that focus, working with two generations of SQL Server simultaneously to root the development in solid, long-standing expertise with SQL Server 2000 and ASP.NET, as well as an early adopter's understanding of SQL Server 2005. 3 Leaf development can be contacted by visiting www.3leaf.com.

2.3 The Baseline Contoso Times Web Application Both the SQL Server 2005 developer and the SQL Server 2000 developer used the same baseline application, which was based on SQL Server 2000 and ASP.NET, as the basis for their development activities associated with this productivity study. The Contoso Times application was a dynamic, data-driven web site that delivered newspaper-style article content to end users.


of 18

http://www.3leaf.com/

2.3.1 Baseline Application Front-End Functionality

The web site was based on an ASP.NET web application. News content was divided into sections that corresponded roughly to traditional newspaper sections, each of which was categorized as one of two types:

•

•

Public content available to all users of the site. This category included the content sections Local News, National News, World News, Baseball, Basketball, and Football.

Premium content available only to registered, authenticated users. This category included the content sections Travel, Entertainment, and Technology.

Registration was free, and the site provided an easy-to-use interface for registering and becoming a member. In addition to enjoying access to premium content, registered users benefited from a recently-viewed-items feature – a special page that listed the last three articles they had viewed. This information was stored persistently so that a user could visit the site months later and still see a list of the last three articles that he or she had viewed.

2.3.2 Baseline Application Back-End Infrastructure

A SQL Server database consisting of a single user-defined table stored all site content. For each article record, it contained an ItemID to uniquely identify the article, a CategoryID to place the article in one of


of 18

the nine content sections, and the text of the title, subtitle, and body of the article. It also included various stored procedures to retrieve, delete, and update article content.

The Contoso Times used forms authentication to identify users. Users who wished to log in could do so through a login form. User names, passwords, and other information regarding registered users was stored securely in a back-end membership database. Once a user had logged in, the Web site used role-based security to differentiate between types of users. Ordinary users could access the site’s public and premium content, but they could not access the site’s administrative pages. Users who were administrators enjoyed access to the administrative portion of the site, where articles could be viewed, edited, and deleted.

2.4 High-Level Overview of Development Phases in the Study This study consisted of the phased rollout of new functionality into the baseline application, much as new features are often added to in-house or commercial applications. In order to provide specific milestones for comparison between development in SQL Server 2005 and SQL Server 2000, the study was divided into five distinct phases, each of which is summarized below.

2.4.1 Phase 1 Functionality: Paid Articles and Subscriptions

The scenario of the first development phase was to fulfill a revenue-generating opportunity for the site. As the popularity of the Contoso Times increased, management decided to charge for some of the content on the site. Adding this functionality required the development staff to build upon the baseline application's ability to manage user accounts, adding secure credit-card payment functionality and the ability to manage payment profiles. Developers also created a secure web service that allows premium subscribers to automatically syndicate content from Contoso Times onto external web sites.

2.4.2 Phase 2 Functionality: Targeted Advertising

Contoso Times requested enhancements in order to offer targeted advertising services to the site's advertising customers, which gave rise to the second development phase of this study. The site functionality associated with this phase was for the AdWords engine to index all Contoso Times stories and match them to relevant paid advertisements, displaying the ads alongside the stories. This provided advertisers with a targeted audience, and it allowed Contoso Times to charge a premium for the service.

2.4.3 Phase 3 Functionality: Middle-Tier Caching

The third phase of the study responded to the need to increase site performance as the readership increased. By implementing middle-tier caching, the developers increased the site's scalability by decreasing the load on the content database. This implementation enabled content to be provided directly from the cache without querying the database, unless the content had changed since it was last cached.

2.4.4 Phase 4 Functionality: Custom and Ad-Hoc Reporting

In order to gauge the performance of the site, as well as the success of the previous changes to the environment, the Contoso Times management needed improved ability to gather metrics associated with the site's revenue-generating activities. The fourth phase of this study addressed these requirements by providing robust reporting mechanisms related to the site. To support this phase, the developers were required to implement a custom reporting solution based on a web-based interface and an ad-hoc mechanism that allowed employees to create reports on an as-needed basis.

2.4.5 Phase 5 Functionality: Offline Smart Client Application for Ad Sales

In the final phase of this productivity study, the developers were called upon to improve the efficiency of the advertising-sales process. Their goal was to achieve that improvement by providing a smart client application that works both online and offline, for use by both tele-sales representatives and field sales personnel. The application must enable sales representatives to generate reports on existing advertisements, as well as to take orders for new advertising sales.


of 18

3.0 Phase 1 Development: Paid Articles and Subscriptions

This phase of development simulated the enablement of a new revenue-generating model for the Contoso Times application by enabling it to charge readers for premium content. A free portion of each premium article (approximately three lines long) is viewable without charge. The pay-for-content model supports three payment modalities:

•

•

•

On Demand: This option allows users to users to pay to view or download content that has been designated with a 'premium' flag on a per-article basis.

Subscription: This option allows users to purchase the rights to view and download all content on the site.

Premium Subscription: This option allows users to purchase rights not only to view and download all content on the site, but also to syndicate it onto their own web sites or other publications, using a web service built for that purpose.

3.1 Technology Requirements Underlying Phase 1 Development A web form-based administration area enables articles to be designated articles as premium content, as well as to create the free teaser portion of paid articles. The capability to flag articles is available only to selected user roles within the Contoso Times organization. A web form is also provided that captures and encrypts user data (using a wrapper class and the Advanced Encryption Standard) such as a username/password pair, contact information, and credit card information. For the subscription model, stored user credentials allow or deny access to premium content.

To facilitate the premium subscription model, the application exposes article content using a web service interface that can be called by the subscriber in order to publish the content on their site. That web service also automatically adds appropriate information to each article that attributes the content to Contoso Times, as well as providing a link back to the Contoso Times site. The web service incorporates authentication mechanisms so that only authorized users can access it.

3.2 Summary of Phase 1 SQL Server 2005 Development

3.2.1 Phase 1 SQL Server 2005 Development Time Requirements

The SQL Server 2005 developer required a total of 27.5 hours to implement the functionality called for in phase 1 of the study. He reported the breakdown of those hours as follows:

Task Time Research 0 hours Manage user setup, authentication, and authorization 6 hours Manage encryption/decryption 8 hours Create web service 8 hours Miscellaneous 5.5 hours Total SQL 2005 Development, Phase 1: 27.5 hours

3.2.2 Phase 1 SQL Server 2005 Platform-Impact Statement

The SQL Server 2005 developer benefited early on from the built-in encryption functionality in the platform. He reported finding it quite simple to use, and based on some Internet research, he was able quickly to put asymmetric encryption into place. Later in the project, the developers refined the project requirements to specify the use of symmetric encryption, in order to more tightly align the development efforts between platforms. Although symmetric encryption was more complex to implement on this platform than asymmetric encryption, he was able to make the change from one modality to the other in


of 18

just two hours. That timing included the required implementation of another level of security, which he fulfilled using certificates.

The developer did encounter a significant development issue during this complex phase of the project, however. He was unable to install Certificate Services on his server in order to issue a certificate to test web service security. While he was able to deploy the security model efficiently, he was unable to fully test it.

The SQL Server 2005 developer realized benefits from the platform's automation with regard to user-account creation and maintenance. The aspnetdb database and login controls simplified operations in this area, leading to distinct advantages in terms of the amount of development time he expended on this phase of the study.



The SQL Server 2000 developer required a total of 98 hours to implement the functionality called for in phase 1 of the study. He reported the breakdown of those hours as follows:

Task Time Research 3.5 hours Manage user setup, authentication, and authorization 43 hours Manage encryption/decryption 30.5 hours Create web service 9 hours Miscellaneous 12 hours Total SQL 2000 Development, Phase 1: 98 hours

This outcome represents approximately 3.6 times the development time required by the SQL 2005 developer for this phase of the study.


The SQL Server 2000 developer began this phase of the study by addressing the need for managing user data with encryption. Since unlike SQL Server 2005, SQL Server 2000 has no built-in encryption functionality, he had to handle encryption outside the database, using .NET cryptographic classes.

One limitation of the chosen approach was that the encryption method returns data as a byte array, which required that the data be stored in binary format, making it non-indexable and non-searchable. Testing and refining the encryption functionality and user registration portion of the application was complicated by the need to handle all data through the encryption class, which added additional time requirements to this phase of the development effort.

3.4 Phase 1 Competitive Outcome The outcome of this phase was that the SQL Server 2005 developer was able to implement the solution in approximately 3.5 developer-days, whereas the SQL Server 2000 developer required slightly more than 12 days.

This difference in development effort is directly attributable to the advances in built-in functionality associated with the SQL Server 2005 environment. By removing the need for developers to manually address issues associated with encryption and other areas, SQL Server 2005 generated very substantial efficiency gains in this phase of the productivity study.


of 18

4.0 Phase 2 Development: Targeted Advertising

This phase of development introduced another revenue-generating feature to the Contoso Times site – the ability to match advertiser's content to corresponding editorial content. On a regular basis, the entire editorial contents of the site is indexed and matched to advertisements so that specific advertisements can be displayed alongside the articles that provide the best match in terms of subject matter.

This targeted advertising mechanism allows the web site to provide various rate structures to advertisers, with an eye toward increasing advertising revenue. For example, advertisers could pay more to have their ads appear more often, or they could cut advertising costs by having their ads appear less often but with only the most relevant stories. This flexibility adds business value for both the advertisers and the site owners.

4.1 Technology Requirements Underlying Phase 2 Development Separate web form-based interfaces allow both new articles and new advertisements to be associated with the system. The advertisement interface supplies rates for having ads appear at various frequencies, collects the advertising content itself, and allows advertisers to enter up to five keywords. When new advertisements are entered into the system, a database trigger launches asynchronous indexing of every article in the system to match the advertisement keywords with target articles in a database mapping table.

A web form collects advertiser billing and contact information, which is stored in an encrypted state. Reporting capability enables advertisers to generate various reports that show click-through statistics on each of their ads, including demographic information about the users who clicked on their ads.




Task Time Research 0 hours Manage encrypted advertiser data, authentication, and authorization 8 hours Manage advertisements, indexing, and tracking 17 hours Develop reporting 5 hours Miscellaneous 2 hours Total SQL 2005 Development, Phase 2: 32 hours


At the outset of this phase of the productivity study, the SQL 2005 developer benefited from the functionality contained in SQL Server Management Studio in creating new tables to handle the advertisements and the associations between the advertisements and the articles. He also made use of the built-in Full-Text Index Wizard in SQL Server 2005 to quickly implement full-text index search on the site's news content, in support of the associations between advertisements and target news content. This functionality also benefited from the fact that one can specify more than one column to index using the wizard in SQL Server 2005, as opposed to only one column in SQL Server 2000.

In terms of the reporting functionality, despite some data errors that required debugging, the SQL Server 2005 developer found overall that working with the reports was straightforward and simple. He also noted that with the Reporting Wizard and Report Designer one has the option to create a consistent overall look


of 18

and feel to the reports rather simply. Overall, he found the integration of the reports to be easy to work with and a source of significant time savings.




Task Time Research 6.5 hours Manage encrypted advertiser data, authentication, and authorization 11 hours Manage advertisements, indexing, and tracking 42.5 hours Develop reporting 15 hours Miscellaneous 10 hours Total SQL 2000 Development, Phase 2: 85 hours



In this phase, the SQL Server 2000 developer again experienced significant delays associated with the lack of automated encryption functionality in SQL Server 2000. As he identified in Phase 1, SQL Server 2000 developers working with encrypted data do not have direct access to the data, such as in Query Analyzer or in Enterprise Manager, since the binary fields cannot be edited. Thus, a quick change that would be simple in a normal database is not possible. A related challenge concerns the fact that every piece of data must be decrypted for reports, which requires the developer to implement a means to handle the decryption.

In order to support Reporting Services 2000 for this part of the study, it was necessary for the developer to use Visual Studio .NET 2003 to design the reports since this is the affiliated product with SQL Server 2000 reporting services, making him unable to benefit from the functionality and interface enhancements in Visual Studio 2005 as he was able to in other portions of the study.

4.4 Phase 2 Competitive Outcome The outcome of this phase was that the SQL Server 2005 developer was able to implement the solution in four developer-days, whereas the SQL Server 2000 developer required slightly more than 10.5 days.

This difference in development time is again attributable to the advantages enjoyed by the SQL Server 2005 developer associated with built-in functionality in SQL Server 2005, particularly with regard to encryption. The SQL Server 2000 may have also encountered some overhead associated with having to switch over to the Visual Studio .NET 2003 environment during part of this phase.

5.0 Phase 3 Development: Middle-Tier Caching

This phase called for the implementation of middle-tier caching to improve performance and scalability of the overall site. By reducing the number of queries against the content database, this implementation reduced the average load on the database server per page-view, resulting in decreased latency from the user's point of view. Because middle-tier caching allows the site owners to support a larger number of users with the same amount of server infrastructure, this strategy enables them to obtain better return on their hardware investment.


of 18

5.1 Technology Requirements Underlying Phase 3 Development This phase implemented notifications when either articles or advertisements change on the site. Those notifications are based on the ASP.NET Cache object, which allows callback functionality when the cached data is invalidated, as well as when dependencies are changed. In the event that an article or advertisement has been added, changed, or deleted, the site cache logic must query the database to obtain current information; otherwise, it delivers the appropriate cached data.




Task Time Research and preparation 3 hours Implement caching 3 hours Total SQL 2005 Development, Phase 3: 6 hours


In this phase of the study, the SQL Server 2005 developer chose to implement caching on the home page, on the summary page when a category is chosen, and on the specific article page. He chose to use the SqlCacheDependency namespace as a means to provide caching reliably and easily. To set up the caching solution on the home page, he programmatically enabled SqlCacheDependency caching. In the other two areas, he set up caching declaratively on the aspx page because of the objectdatasource that was present in both pages.

The SQL Server 2005 developer did have to devote some extra development effort to declarative caching, since by default, it would cache everything on the page. That characteristic would not allow for dynamic rendering of data in some circumstances, so he had to manage caching programmatically to be more specific about what would be cached and when. He reported, however, that this implementation was quite simple to implement overall.




Task Time Research and preparation 3.5 hours Implement caching 3 hours Total SQL 2000 Development, Phase 3: 6.5 hours



The SQL Server 2000 developer reports that he enabled caching on the SqlDataSource or ObjectDataSource controls on each web page, creating a dependency on a given SQL Server table. When the control retrieves data, it caches it, and only revisits the database if data in that table has changed.


of 18

His solution also executes the aspnet_regsql.exe application, which prepares SQL Server for handling the notification relationship with ASP.NET, announcing when tables have changed.

5.4 Phase 3 Competitive Outcome Both the SQL Server 2005 solution and the SQL Server 2000 solutions took less than one developer day to implement for this phase of the productivity study. Since the functionality in this portion of the study is predominately based on the feature set found in ASP.NET 2.0, the developers faced the same issues and had access to the same tools for the development of middle-tier caching functionality. Thus, this phase showed little difference between the two development environments, even with their slightly different approaches to solving the problem.

Since both developers implemented a high-level solution that primarily used features of ASP.NET 2.0 rather than features specific to SQL Server 2005 or SQL Server 2000, there was little difference in the development effort required for this phase. Both benefited from the ability of ASP.NET 2.0 to work in conjunction with SQL Server 7.0, 2000, and 2005 to handle caching in a mostly declarative way. This capability leveled the playing field between the two development environments.

6.0 Phase 4 Development: Custom and Ad-Hoc Reporting

This phase called for the addition of reporting mechanisms to gather metrics related to the success of various aspects of the site's operation, such as subscription information, advertising impact, and paid article consumption. It also allows members of the Contoso Times staff to generate ad hoc reports on subscription and advertising data. Such specialized reports can be valuable, for example, in support of sales activities. By empowering staff members to create reports based on their individual needs, when they need them, those staff members are able to perform their jobs more efficiently, and the costs associated with designing custom queries for one-time reports are avoided.

Ad hoc reports are also available to advertisers, providing post-sales support at very low cost to the Contoso Times organization by enabling advertisers to independently quantify the success of their advertising investment.

6.1 Technology Requirements Underlying Phase 4 Development Custom Reports are available only to members of the Contoso Times staff, while ad hoc reports are available to advertisers as well. Users can save ad hoc reports for future use.

The site implements custom reports using a web-based interface that is available to all members of the Contoso Times organization, although certain reports are available only to certain members of the company. The reports themselves are static and printable, incorporating no 'drill-down' or dynamic sorting functionality. Reporting users are based on Contoso Times user accounts with temporary assignment of higher access privileges for the purpose of running the report.




Task Time Research and preparation 4.25 hours Build reports 10.25 hours Miscellaneous 2 hours Total SQL 2005 Development, Phase 4: 16.5 hours


of 18


The SQL Server 2005 developer reported that he found the built-in Report Viewer control and the new Report Builder in Reporting Services quite helpful in creating local reports and ad hoc reports, respectively. In a real-world deployment, moreover, SQL Server 2005's new Report Builder would provide a user-friendly business-query model that would enable general users to generate additional ad hoc reports. Report Builder is a ClickOnce Windows Forms application launched from the Report Server that enabled the SQL Server 2005 developer to build the required reports more rapidly than would otherwise have been possible.




Task Time Research and preparation 10 hours Build reports 19.5 hours Miscellaneous 0 hours Total SQL 2000 Development, Phase 4: 29.5 hours



The SQL Server 2000 developer had to create local reports and ad hoc reports as a Report Project using Reporting Services. This approach required more steps to develop and deploy that the equivalent effort by the SQL Server 2005 developer, who benefited from the use of Report Builder, which is not available in SQL Server 2000. Moreover, in SQL Server 2000, ad hoc reporting had to be developed as a custom solution, and the SQL Server 2000 developer was also compelled to install Visual Studio .NET 2003 in addition to Visual Studio 2005, specifically to support SQL Server 2000 Reporting Services.

6.4 Phase 4 Competitive Outcome The outcome of this phase was that the SQL Server 2005 developer was able to implement the solution in slightly more than two developer-days, whereas it took the SQL Server 2000 developer approximately three and a half days.

The most significant contributor to this difference was the time required to create the reports themselves, which took the SQL Server 2000 developer nearly twice as long as the SQL Server 2005 developer. As mentioned in the platform-impact statements above, this difference was attributable in large part to the addition of Report Builder in SQL Server 2005. This type of advance in the automation of common, tedious tasks enables developers and administrators to complete such tasks more quickly, allowing them to focus more on value-added responsibilities.

7.0 Phase 5 Development: Offline Smart Client Application for Ad Sales

This phase of the site development targeted improving the efficiency with which sales personnel can sell advertising by providing a smart client application that works both online and offline. The application incorporates two general spheres of functionality:

• Reporting, Charting, Graphs, and Research: This aspect of the application primarily targets pre-sales support by enabling sales representatives to collect and report on advertising and viewing data.


of 18

• Order Fulfillment: This portion of the application lets sales staff actually take orders and enter the relevant information to close the sale. It supports both the advertising information itself, as well as the contact and payment information of the purchaser.

7.1 Technology Requirements Underlying Phase 5 Development The reporting, charting, graphs, and research portion of the application takes advantage of the reporting functionality already implemented in Phase 4 of the study, but it also incorporates the local caching and synchronization functionality to allow this reporting to take place while the user is offline. It uses reporting functionality built into Visual Studio 2005 to create ad hoc reports from the local data set, which is stored using SQL Server Express.

The order-fulfillment part of the application provides a Windows Forms interface that encrypts personal data including contact and credit card information using CLR classes to implement the AES 192-bit algorithm. Windows Forms validation was used to enforce correct formats for the corresponding fields. Merge replication uploads the data to the central server, using partitioned articles to limit the data returned to the representative. A simplified deployment model uses ClickOnce technology to enable sales representatives to easily receive updated versions of the application via web download from a virtual directory on the Contoso Times web site. This flexibility enables sales representatives to move about freely and perform their duties in an efficient manner with or without connectivity to the network, while also letting site administrators govern their access to data.




Task Time Research 0 hours Create interface and advertiser/advertisement data management 8 hours Implement database access and replication 2 hours Implement reporting 4.5 hours Miscellaneous .5 hours Total SQL 2005 Development, Phase 5: 15 hours


During the development of the smart client application, the SQL Server 2005 developer reported substantial efficiency benefits in setting up replication. He set up both the publication and subscription engines to ensure that they would replicate the data using built-in wizards. He used the 'transactional publication with updatable subscription' functionality, which didn’t exist in SQL Server 2000. This type of replication enables changes from the publisher to stream to the subscriber after the initial snapshot, and changes from the subscriber are applied at the publisher.

At one point, he had to make changes to the subscription, which required him to drop the subscription and re-create it, which would have been quite time-consuming in SQL Server 2000. In SQL Server 2005, he was able to simply step through a wizard and change the update process from occurring on a 'simultaneously commit changes' basis to occurring on a 'queue changes and commit when possible' basis. The SQL Server 2005 developer stated that this made implementing the offline/online capability of the application very easy.


of 18




Task Time Research 8 hours Create interface and advertiser/advertisement data management 19 hours Implement database access and replication 6.5 hours Implement reporting 15 hours Miscellaneous 1.5 hours Total SQL 2000 Development, Phase 5: 50 hours



The first significant limitation faced by the SQL Server 2000 developer in this phase related to the lack of built-in encryption functionality in SQL Server 2000, as he had encountered in Phase 1. In order to encrypt the advertiser personal information, he had to handle encryption outside the database, using .NET cryptographic classes. Because of this approach, the developer was unable to bind directly to the text boxes, drop-down lists, and other customer-data controls, which introduced significant added development time to this phase of the project.

Another shortcoming of being compelled to use the ItemEncryptor class was the developer's resultant inability to take advantage of the ability to establish a data source representing selected tables, views, stored procedures, and user-defined functions from the database. He was unable simply to drag selected columns or an entire table directly onto the form, allowing Visual Studio to create controls directly bound to the data source. Instead, he had to accomplish data binding by populating each control through code, which also added significantly to development time.

7.4 Phase 5 Competitive Outcome The outcome of this phase was that the SQL Server 2005 developer was able to implement the solution in just under two developer-days, whereas it took the SQL Server 2000 developer approximately six days.

This differential was attributable to functionality added in SQL Server to automate tasks. For example, the SQL Server 2005 developer reported substantial time savings due to the availability of wizards in SQL Server 2005 that allowed him to rapidly set up replication and carry out changes to the subscription. This simplification of these tasks is a good example of the advantages that are available from the added maturity of the SQL Server 2005 environment, relative to SQL Server 2000.

The addition of built-in encryption in SQL Server 2005 again played a large role in simplifying development of the solution. By adding commonly used functionality such as this to the SQL Server platform, developers are able to more efficiently build features into their applications, with resultant savings in development efficiency and faster time to deployment.

8.0 Overall Outcome

For the total development time of all five phases of the study, the SQL Server 2005 developer reported 97 hours, or just over 12 developer-days. The SQL Server 2000 developer reported total development time of 270 hours, or just under 34 developer-days, approximately 2.8 times the amount of time required by


of 18

the SQL Server 2005 developer. The comparison of the hours required by each developer is summarized in the chart below:

0

50

100

150

200

250

300

Developer Time

(hours)

SQL 2005 SQL2000

Phase 5Phase 4Phase 3Phase 2Phase 1

This difference in the resources required by the two developers to create the solutions under SQL Server 2005 and SQL Server 2000 reflects the increased value and programmability of the newer database platform.

9.0 Key SQL Server 2005 Usability Benefits Reported by Developers

The developers involved in this study reported significant advantages to the SQL Server 2005 environment in day-to-day usability, relative to the SQL Server 2000 environment. The study found that the combination of SQL Server 2005 and Visual Studio 2005 provides significant efficiency advantages over the earlier platform1. Following are a few of the key usability differences between the two platforms, as they pertain to this study.

9.1 Encryption Advantages As mentioned in the discussions above, the SQL Server 2005 developer reported finding the built-in encryption quite useful and easy to implement. Absent this functionality, the SQL Server 2000 developer had to implement encryption programmatically, which required code to encrypt and decrypt sensitive data every time he needed to access it. Conversely, by having this capability on the database side, the SQL Server 2005 developer was able to write it once in a stored procedure and reuse it as needed, making changes as needed without recompiling the Contoso Times application.

9.2 Reporting Advantages As mentioned in the Phase 4 discussion above, the SQL Server 2005 developer benefited from the built-in Report Viewer control and the new Report Builder in Reporting Services in creating local reports and ad hoc reports, respectively. The SQL Server 2000 developer had to build this functionality as a Report Project using Reporting Services, which required more steps to develop and deploy. Moreover, in SQL Server 2000, ad hoc reporting had to be developed as a custom solution.

9.3 Replication Advantages As mentioned in the Phase 5 discussion above, the SQL Server 2005 developer benefited substantially in terms of efficiency when he set up replication. He used built-in wizards to set up both the publication and subscription engines to ensure that they would replicate the data. He indicated that this automated 1 For details about the hardware and software platforms used by each developer, see the appendix of this document.


of 18

functionality substantially decreased the time required to implement the offline/online capability of the application.

9.4 Interface and File-Management Advantages The SQL Server 2005 developer responded especially favorably to the addition of SQL Server Management Studio in the new platform. By incorporating functionality into a single tool that had been spread among two or three tools in SQL Server 2000, the developer realized distinct efficiency gains. In particular, he benefited from being able to test queries and resolve debug issues without switching back and forth between multiple environments. Visual Studio 2005's inclusion of a local web server also provided significant advantages. This functionality simplified development by removing the need to configure Internet Information Services. Moreover, it enabled him to put web applications in any directory he wanted, moving them at will with a simple cut and paste operation.

9.5 Common Language Runtime Integration Advantages The CLR integration in SQL Server 2005 proved invaluable in the creation of the AdWord trigger in Phase 2 of the study. The SQL Server 2005 developer was able to create a trigger that leveraged a great deal of programmatic logic that was not available to the SQL Server 2000 developer. Additionally, to install the trigger on the database, it required only that the SQL Server 2005 developer deploy the triggers, and the environment accomplished the rest of the process transparently, which included dropping and recreating the trigger on the database every time the solution was built and run.

10.0 Summary and Conclusion

The findings of this study demonstrate the real efficiency benefits that are available to developers in their day-to-day activities by upgrading from SQL Server 2000 to SQL Server 2005. The comparative 97 hours required by the SQL Server 2005 developer to perform the same tasks that took 270 hours for the SQL Server 2000 developer are emblematic of the unprecedented value and functionality afforded by the newer platform.

The addition of built-in functionality such as encryption and user authentication, the increased integration with Visual Studio and the CLR, and automation of various common tasks were instrumental to the efficiency gains realized by the SQL Server 2005 developer in this study. Increased programmability, efficiency, and value are available for business-critical implementation by enterprise developers today.

Learn more about the benefits of SQL Server 2005 at http://www.microsoft.com/sql/default.mspx.


of 18

http://www.microsoft.com/sql/default.mspx

Appendix: The Development Environments

The developers were chosen specifically for their extensive and similar levels of expertise in relation to the tasks required in this productivity study. Both have seven years of experience programming with various generations of SQL Server, as well as seven to eight years of experience with web development, including ASP and ASP.NET. Likewise, both have worked closely with Windows Server platforms for seven to eight years. Overall, they brought equivalent skill sets to the study, in terms of the core technologies and tools used.

Both developers used similar environments, consisting primarily of their respective SQL Server versions and Visual Studio 2005. In addition, the SQL Server 2000 developer used Visual Studio .NET 2003, which was required to support SQL Server 2000 Reporting Services.

The development process focused as much as possible on the use of SQL Server tools, as opposed to tools provided by Visual Studio. For example, in Phase 5, the developers were compelled to use merge replication with SQL Express, as opposed to building out a custom synchronization mechanism for the smart client application.

SQL Server 2005 Software Environment Database Environment: SQL Server 2005 CTP June (moving to CTP Sept)

IDE: Visual Studio 2005 Beta 2 (moving to RC1 as the study progressed)

SQL Server 2000 Software Environment Database Environment: SQL Server 2000

IDE: Visual Studio 2005, Visual Studio .NET 2003 (required by SQL Server 2000 Reporting Services)

This productivity study was conducted by a 3 Leaf, a firm unaffiliated with Microsoft. Since 2000, 3 Leaf has specialized in creating reference, proof of concept, and sample applications, targeted training content, and demonstrations and keynote presentations for Microsoft, Intel and others, in addition to more traditional development efforts. Its developers are early adopters of a wide range of technologies.

For more information about 3 Leaf, please visit www.3leaf.com.

To contact 3 Leaf directly, please email [email protected].


of 18

http://www.3leaf.com/

Comparing Enterprise Development...

Documents

Transcript of Comparing Enterprise Development...