Comparison of database management systems

¶ … Management Systems

Appendix a Project Process Integration Diagram

Appendix B Process Flow Chart

Open Source Security Tools

DBMS Features Analysis

Figure 2 Three Pillars of Security

The database market continues to consolidate with Oracle acquiring Sun Microsystems, who had acquired MySQL AB in January, 2008. While there are dozens of vendors in the database management systems (DBMS) market, IBM, Microsoft and Oracle dominate the market with 88% market share according to International Data Corporation (Worthen, Kiviniemi, 2009). As DBMSs are often customized to the specific needs of the companies using them, this consolidation does not spell the end of the heterogeneous nature of this market. Organizations' needs are so diverse and the structure of their data and use of applications so varied, that the many vendors in this market are staying relevant and growing as a result. This fragmentation of the database market has opened up significant opportunities for Computer Associates (CA), Software AG, Sybase and many open source databases the highest performing being Ingres, MySQL and PostgreSQL. The lower Total Cost of Ownership (TCO) and speed of new product development innovation in open source, specifically MySQL, has redefined the direction of the DBMS market globally (Hyatt, 2008). Oracle as a result continues to be in a defensive posture with regard to the economics of open source, hence their acquisition of Sun Microsystems which includes MySQL AB (Worthen, Kiviniemi, 2009). As a result of these market dynamics, MySQL continues to excel from a technological and user adoption perspective (Denton, Peace, 2003). The intent of this analysis is to evaluate MySQL's technical functionality vs. To PostgreSQL, Oracle and Microsoft SQL Server. Components of their current offering, market presence and strategy are defined.

Introduction

For the typical enterprise today the need for integrating a variety of databases is critical, as each contributes a unique strength to an enterprise-wide strategy. The majority of enterprise and organizations are integrating several databases together, each specifically designed for a given applications' needs (Giacomo, 2005). Open source DBMS have had the advantage of a lower TCO, faster potential time-to-market given the open source community's level of innovation relative to enterprise software companies (Hyatt, 2008) and the expanded security model which MySQL has been able to fine-tune the shared security model of their kernel as well (Pereira, Muppavarapu, Chung, 2006). In analyzing a performance benchmarks of open source vs. enterprise applications (Denton, Peace, 2003) (Giacomo, 2005) (Hyatt, 2008) (Pereira, Muppavarapu, Chung, 2006) and (Ramsey, 2003) it is clear that the performance of MySQL on platform scalability and user adoption is exceptional compared to other competing DBMS. MySQL also has concentrated on adoption, and has a result significantly driven up the total user count on their systems as a result (Hyatt, 2008). Application data integration, 3rd party application support, performance analysis, scalability, database programming and application development are all critical factors as well. This report concentrates on a compatibility and performance of Microsoft SQL Server, MySQL, Oracle 11g and PostgreSQL. The specific product versions include Microsoft SQL Server 2008, MySQL 5.1, Oracle 11g, and PostgreSQL 8.3. The majority of these databases were introduced in the late 2008 timeframe, with Oracle 11g being released in October, 2007. The three dominant uses of databases are Online Transaction processing (OLTP), data warehouses and specialized databases. Each of these databases began with the explicit design goal of providing relational database performance in the context of OLTP in the context of role-based workflows. The catalyst of growth for databases, both closed source and open source, continues to be the growth of roles-based technologies including Web Services and Service-Oriented Architectures (Denton, Peace, 2003). MySQL and PostgreSQL have historically lagged in terms of enterprise application support yet have had exceptional growth in security modeling and Web Services. For a matrix overview of representative security utilities and tools on open source DMBS including MySQL please see the Appendix C Open Source Security Tools in the Appendix. In conjunction with the security models open source DBMS developers are adopting, analysis indicates open source DMBS can be more security that closed source (Pereira, Muppavarapu, Chung, 2006) (Ramsey, 2003). These configurations of DBMS are highly dependent on the security requirements and security audit procedures of the organizations adopting them (Pereira, Muppavarapu, Chung, 2006). From the analysis completed for this paper the scalability and platform support for open source, as of 2006, continues to surpass closed source. As the TCO is much lower for MySQL and PostgreSQL there has also been a corresponding increase in interest from system integrators and resellers in these open source DBMS. The reason is that the maintenance frees for Oracle 11g and Microsoft SQL Server continue to escalate rapidly which has also created interest in closed database customers' organizations to seek out the cost advantages of open source as well (Spanbauer, 2008).

Discussion

Analyzing attributes or characteristics of the current offerings of closed database vendors Microsoft SQL Server 2008 and Oracle 11g in addition to open source vendors MySQL and PostgreSQL is the intent of this analysis. Within the current offering, the criteria of performance and scalability, security and adoption of security models, application development, database security, platform support and database administration can be used as a common set of criteria for evaluating closed source and open source offerings. Next, the strategy of the vendors including their product strategy direction, and commitment to the specific design objectives are discussed. Finally factors including company financials, adoption rate, training and certification and system integrator and partner support are evaluated. What emerges from this analysis is an assessment of each of the DBMS.

Across the four DBMS being included in this analysis, it is noteworthy that MySQL 5.1 has the greatest level of reliability as shown by the multiple technologies that are used. Even Oracle 11g, the highest capitalized DBMS from a product development standpoint, has only bitmapped, B-tree, IOT and function-based security. It does not however include full-text, which is essential when a database is being as the foundation of an enterprise content management system for example (Pereira, Muppavarapu, Chung, 2006). Table 1, DBMS Features Analysis provides an assessment of these applications across the functional areas of data storage, indexes support, data integrity, replication, interface methods and Advanced Features. Of the open sources databases the one with the best possible scalability is MySQL. This is specifically due to the design of the OLTP engine and structure within the applications' platform (Ramsey, 2003) (Spanbauer, 2008).

Figure 1: DBMS Features Analysis

MySQL 5.1

PostgreSQL 8.3

Oracle 11g

Microsoft SQL Server 2008

Data Storage

Reliability

MyISAM, InnoDB, Berkeley DB, full-text

Postgres

Bitmapped, B-tree, IOT, function-based

Clustered, nonclustered

Scalability

Large/Very Large

Large

Large/Very Large

Large/Very Large

Indexes

Single and Multi-column, primary key, full text

Yes

Yes

Yes

Yes

Data Integrity

ACID compliance, row-level locking, hot backup and partial backup

Yes

Yes

Yes

Yes

Replication

Single master

Yes

Yes

Yes

Yes

Multimaster

Yes

Yes

Yes

Yes

Clustering

Yes

Yes

Yes

Yes

Interface Methods

ODBC/JDBC, C/C++, & J2EE

Yes

Yes

Yes

Yes

Advanced Features

Stored procedures, views, triggers, sequences and cursors

Yes

Yes

Yes

Yes

Sources: (Denton, Peace, 2003)(Giacomo, 2005)(Goans, Leach, Vogel, 2006) (Hyatt, 2008) (Ramsey, 2003) (Pereira, Muppavarapu, Chung, 2006) and websites

Each of the four database are now analyzed according to the nine attributes of platform scalability, user adoption, application data integration, 3rd party application support, performance analysis, database programming and application development. While the closed source vendors have gone aggressively in the direction of creating Service Oriented Architectures (SOA) and composite applications through Web Services (Pereira, Muppavarapu, Chung, 2006) open source DBMS providers have continued to focus on making their core OLTP performance improve significantly over time (Hyatt, 2008).

As a result of the open source DBMS providers concentrating on this aspect first their product roadmaps have taken them in the direction of supporting greater platform breadth which according to studies from AMR Research, Gartner Group and International Data Corporation (Hyatt, 2008) they are achieving greater levels of RDBS and associated application scalability as well. In analyzing the four vendors including in this paper, MySQL is the most scalable from both an DBMS and application platform standpoint (Pereira, Muppavarapu, Chung, 2006). This is due to the kernel for MySQL being scalable and also modular enough for continual improvement and addition of functionality. All of these factors in conjunction with their support for Very Large Database (VLDB) architectures has open source vendors delivering greater levels of scalability and also more agnostic or non-vendor specific as well. Microsoft and Oracle both tend to see integration often from a purely myopic view; they see only the integration of their components toe ach other and to partners' applications. Open source is forced due to its heavy reliance on innovation to concentrate on having a more heterogeneous and therefore varied integration strategy (Hyatt, 2008). This forces DBMS vendors to continually benchmark OLTP, data warehousing, and Web-based application development to ensure they are continually improving inter-database and inter-process scalability and performance (Pereira, Muppavarapu, Chung, 2006). This has become a best practices specifically for open source DBMS vendors as they seek to differentiate through the use of greater levels of integration (Goans, Leach, Vogel, 2006). Another factor in open source database vendors leading this area is the focus on multiple models for scalability as can be seen in Table 1: DBMS Features Analysis. The support of multiple models of scalability also concentrates on OLTP transaction integration specifically within roles as defined in the security model (Pereira, Muppavarapu, Chung, 2006). Security will be discussed later in this analysis.

The second factor used to evaluate closed source vs. open source DBMS are the application development environments that each has. Closed source vendors have turned application development environments into revenue-producing lines of business (Worthen, Kiviniemi, 2009). As a result, closed database vendors have a slight edge in the areas of application development. There are also the initiatives of Service Oriented Architectures (SOA) and distributed application use by closed database vendors Microsoft and Oracle. The integration and use of application development is also a key catalyst for closed source vendors to ensure their maintenance fees can also be sustained over time (Hyatt, 2008). Open source DBMS vendors and providers have benefitted from the broad community of developers, yet it is not as focused on profitability as the primary goal; it is focused on breadth of OLTP engine support and Application Programming Interfaces (API) to ensure broad adoption (Ricadela, 2009). The pervasiveness of these open source extensions rivals closed yet, yet the latter is more aligned and attuned to supporting maintenance cycles and contract management of customers' contracts. In fact the reliance on contract management and the recurring revenue model of closed source DBMS vendors is the primary source of their most profitable revenue over time (Denton, Peace, 2003). This factor is the catalyst for their heavy levels of investment in application development environments and platforms on the part of Microsoft and Oracle especially. Oracle's Fusion initiative is an SOA platform that integrates role-based applications published as Web Services over an XML network (Spanbauer, 2008). Microsoft's .NET architecture is also specifically designed to provide for scalable and role-based Web Services across non-Microsoft systems. Fusion and .NET are the closed source DBMS vendors' approaches to creating a role-based business process management (BPM) platform. Due to these factors, application development is considered more enterprise in scope for closed source vendors.

Database security as a criterion for evaluating the closed source vs. open source vendors in this analysis illustrates how effective innovation of a DBMS can be when there are high levels of collaboration in a user community. Microsoft and Oracle both have defined role separation, reporting and data availability as part of their security models (Pereira, Muppavarapu, Chung, 2006) from a closed database standpoint. Yet the open source vendors included in this analysis have a more thorough interpretation of the three pillars of an effective security strategy (Kamel, 2009). These three pillars include creating a more consistent and uniform approach to path management and discovery & classification of risks, the development of consistent preventative measures for encryption, data masking and change management as well. Lastly and most significant is the need for defining intrusion detection, auditing, monitoring and vulnerability assessment (Kamel, 2009). Taken together these are the factors that define a scalable security strategy over time. On the first area of path management and discovery & classification, open source vendors have devised authentication and authorization logic to the role level for the DBMS platforms, while close source vendors rely on ancillary security applications for this (Kamel, 2009). Just as the basic licensing model of open source DBMS lends itself better to TCO advantages, this integration of security into the foundation of an open source DBMS provides a significant cost advantage (Spanbauer, 2008). The second aspect of security platform differences between open source and closed source are the preventative measures taken to provide higher levels of encryption while reducing the risk of data masking. All of these factors are critical for change management to be more effectively managed at the database, process and organizational levels (Kamel, 2009). Appendix C: Open Source Security Tools provides examples of the level of innovation being achieved in the development of security (Hyatt, 2008). The last column or tier of an enterprise-wide DBMS security strategy is the development of auditing, monitoring and threat assessment methodologies (Iyer, 2009). In this specific area Microsoft SQL Server 2008 and oracle 11g have defined scripts that can automate the auditing process and also launch audits of each database entry points randomly to test the scalability of security over time (Kamel, 2009). Audit data is then analyzed and provided within a dashboard that Data Base Administrators (DBAs) use to evaluate overall security and ACID-compliance as is seen in Table 1: DMBS Features Analysis. Using the series of analyses (Iyer, 2009) (Kamel, 2009) Table 2: Three Pillars of Enterprise Database Security illustrates how the three specific pillars of security interrelate and support role separation, reporting and data availability. Across this three pillars mode open source DMBS data structures have been shown to provide a greater level of data integration and support (Iyer, 2009).

Figure 2 Three Pillars of Security

Sources: (Iyer, 2009) (Kamel, 2009) (Giacomo, 2005) (Pereira, Muppavarapu, Chung, 2006) (Ramsey, 2003)

The next two criterion that are used to evaluate closed source and open source DBMS systems is the direction of the product strategies, commitment to a given architectural direction and the differing strategies in pricing. Across these three factors, the most prevalent in terms of its impact on the industry is the open source model of pricing to a single license and not necessarily requiring maintenance and service contracts as closed source DBMS providers often require in order to get software updates. Maintenance charges for Oracle and Microsoft are often up to 22% of the purchase price of the application, translating into millions of dollars per year in just support costs (Kamel, 2009). Due to this fact closed source DBMS are often part of Capital Expense (CAPEX) budgeting within a company (Iyer, 2009). CAPEX investments by definition need to go in front of the Board of Directors for approval. This significantly slows down adoption. The use of Operating Expense (OPEX) pricing strategies by open source DBMS vendors has led to significant market share gains over time and higher levels of adoption. The economics of this pricing strategy has also had a signification effect on open source DBMS vendor's financials (Pereira, Muppavarapu, Chung, 2006) (Spanbauer, 2008). As a result of the ability to expense investments in open source DBMS, the financial strength of companies including MySQL, as a division of Oracle now due to the acquisition, are superior to closed source vendors. CAPEX requires a longer sales cycle and more focus on the Return on Investment (ROI) of the investment in the DBMS.