Last April, IBM and MySQL AB announced an agreement to port the open-source database called MySQL to the IBM System i platform. This announcement was made at the MySQL Conference & Expo 2007, held at the Santa Clara Convention Center just days before COMMON in Anaheim.
Curious Initial Customer Reactions
Some COMMON attendees reacted to the announcement with a mixture of excitement and mild incredulity. After all, isn't DB2 the database of preference on the System i? Isn't MySQL one of those "buzzword" technologies that—when push comes to shove—doesn't cut the mustard in terms of enterprise-level security and resilience?
Moreover, in some respects, the announcement at COMMON looked like a repeat of the previous year's event, when IBM GM Mark Shearer revealed that Zend was porting its PHP services to the System i. In fact, a number of cynical attendees privately asked, "What's Mark going to announce at the next year's COMMON? Microsoft Access on the System i?"
A One-Two Punch for Application Availability
But the one-two punch of PHP and MySQL, announced in two consecutive years, does, in fact, confirm an ongoing strategy by IBM to lower the threshold for application development on the System i. And, more importantly, it heralds an increase in the number of potential applications that can run on the platform, making the platform more relevant to the software industry at large.
Popularity, Popularity, Popularity!
To appreciate this strategy as a layperson, one need only Google "MySQL PHP." Now, compare the results of that search with a similar search for "i5 DB2." The first query delivers more than 158 million Web pages related to MySQL PHP, while the second search (so much closer to our traditionalist hearts) delivers a paltry 500 thousand. (Searching for "OS/400 DB2" delivers even fewer pages: a mere 472 thousand. Further combined searches with words like "RPG AS/400" pares the number of returned pages yet further.)
The cynics among us will note that this is far from a scientific means to determine the actual importance of the System i's database to the business world. However, it does point to a discrepancy in public market exposure. It indicates that popular mindshare certainly rates MySQL and PHP significantly higher than our favored application services. This "public awareness discrepancy" heightens our understanding of IBM's dilemma for the System i, especially when we considers the histories of all these products.
For instance, MySQL has been around since 1986, but PHP is still in its infancy, created in 1995. By comparison, i5/0S and i5 DB2 have been around significantly longer (though branded with other monikers) yet still struggle for public awareness.
To System i traditionalists, this may cast MySQL and PHP into the roles of "those @#$% upstarts"! But to the rest of the world, MySQL and PHP on the System i create a "me too!" image of the platform and expand the credibility of the i5/OS significantly.
What's a Manager to Do?
This leads us to a number of interesting questions:
- How important will these two announcements (PHP and MySQL) be to the current System i customer base?
- How will the delivery of these services impact that actual IT System i application infrastructure?
- Should we, as IT managers and application managers, begin investigating that potential?
- Will the announcements really expand the use of the System i or even the number of systems that IBM can sell?
Importance to Customer Base
Regardless of our personal preferences for one computing platform over another, the reality of today's IT world is very simple: The cost of new software is the single most important factor driving the decision upon which a computing platform is chosen.
Consider how IT must rationalize its choice of computing platforms to its management. Management really has difficulty telling us the functional differences between a System i and an iPod, much less the difference between DB2 and SQL, or RPG and PHP. All management knows for certain is that the timeline for creating new software is always too long, while the cost of buying new software is always too high. That's management's area of expertise, and very often it drives traditionalists in the System i IT department crazy! But it's reality.
Reducing Costs and Timelines
The PHP/MySQL combo helps System i IT significantly remove the cost impediments for management. Though the licensing of these open-source services for the System i are certainly not free, the services themselves open a door to literally thousands of open-source applications that are essentially free—or at least significantly lower in cost than comparable applications from the traditional System i or Microsoft or UNIX vendor community.
Moreover, PHP is a rapid application development (RAD) scripting language that makes traditional commercial-grade middleware services pale by comparison. Consider that you can take an open-source application written in PHP and modify it to meet business goals relatively quickly, reducing in-house costs of development significantly. Try doing something similar with WebSphere applications or Microsoft applications, and you have an argument that definitely gets management's attention.
Finally, look at the number of consultants who are marketing their skills in PHP/MySQL—along with their hourly rates—and suddenly the penny will drop: Management can get more bang for the buck with a PHP/MySQL application than with similarly functioning commercial applications written by name-brand vendors. It's a formula that will warm the cockles of most small-to-medium-sized organizations.
Impacting the Application Infrastructure?
The larger technical questions that are on many System i IT managers' minds relate to how MySQL/PHP applications impact the overall responsibility of IT. There are multiple concerns:
- Complexity in a mixed application environment
- Application lifecycle
Each of these concerns is dependent upon the particular application that is being created or purchased, and there is no panacea.
Complexity within any particular application can range from "simple, stand-alone function" to "highly integrated, cross-platform, distributed mission-critical suite." For instance, if a particular PHP/MySQL application must be highly integrated to pre-existing back-end RPG/DB2 applications, levels of complexity must be analyzed before a decision on that application can be cost-justified.
The good news is that IBM plans to embed the MySQL open-source database into the integrated i5/DB2 database, which should, theoretically, enable easy cross-language access.
It's still too early to tell how this tactic will impact the performance of the System i, but given the IBM track record for optimizing System i performance, we believe IBM will adjust its operating system accordingly to deliver adequate performance. Certainly, scalability will not be an issue.
Manpower and Application Lifecycle
On the other hand, maintenance, application lifecycle, and training concerns hinge upon personnel issues. For instance, what happens if the key programmers with the knowledge of PHP and MySQL move on? How will IT find replacements?
This is not a minor concern, based upon real-life IT experience with ported services. For instance, IBM and its vendors sometimes promote unique proprietary application modernization technologies. Then, as years go by, these players shift strategies or technologies, leave the market, or consolidate, leaving IT with no trained resources and no way to service the applications that were created.
Fortunately, this does not appear to be a potential problem with today's PHP/MySQL environment. Literally thousands of resources are available for hire or for contract. PHP and MySQL classes are taught widely at most post-secondary educational institutions. Finally, if IT makes the effort to choose open-source applications or to include the cost of the source code in its purchase price, it will have secured its ability to control the lifecycle of each product purchased.
The final concern is with the basic security of PHP scripts and MySQL access, and indeed there have been concerns with the vulnerabilities of these Web-facing services in the past. That being said, the popularity of these services—combined with the massive numbers of developers who use them—has historically proven the responsiveness and the resiliency of the open-source community to solve discovered vulnerabilities.
If you evaluate the responsiveness of this combined community to its security threats and then compare Microsoft's record of responsiveness to similar threats, you get a clear picture of how the open-source paradigm can actually benefit the security of organization: Open-source middleware services tend to have more developers dedicated to problem-solving than commercial middleware services do.
Finally, because MySQL and PHP on the System i reside inside the larger security envelope of i5/OS, there's a general belief—still not proven—that there is a stronger potential to keep hackers at bay.
Should System i Managers Invest the Time to Investigate?
The simple answer is yes! PHP and MySQL books, tutorials, and sample applications abound on the Internet. Learning the basics takes time, but it expands the horizons and the potential of both your developers and your users.
Zend, the creator of PHP, offers a trial IDE called Zend Studio that helps with the learning curve of the development cycle. But there are other IDEs as well.
By the same token, MySQL is available through MySQL AB or from other open-source sites. These services can be executed on a personal computer at no cost for the software. Later, when IT's strategic decision has been reached, the basic skills learned can be readily transferred to the System i implementation of these services.
Will These Services on the System i Sell Boxes?
IBM is making the strategic bet that they will! However, the jury is still out, and tactical implementation of the PHP/MySQL strategy is still working its way through IBM's Business Partner channels. The MySQL announcement is only a month old, and no real System i version has been actually delivered to date. PHP, by comparison, has been out for a year now, and while downloads of the Zend product have been robust, the actual impact on new System i footprints seems minimal. (System i sales were down last quarter.)
The unfortunate reality of IBM's Business Partner structure is the System i's unique value statement. This difficulty is often enhanced when the Business Partner is also selling competing IBM boxes. In this light, the PHP/MySQL strategy only seems to remove the arguments against the purchase of a System i. The perception by customers seems to be "Why should I buy a System i to run a PHP/MySQL application if I can get the same application to run on a PC server?" Cost, again, is the sticking point.
But IBM's recent announcement of the low-to-mid-priced System i models addresses these concerns. Still, once again, it appears IBM is attempting to remove the arguments against a System i purchase. Bit by bit, Rochester is chipping away at these sales obstacles, believing that the basic value statements of the computing platform will erode the customers' excuses and prejudices.
Still, the underlying question remains: Is this a marketing strategy? Or is it merely a survival strategy to keep the platform relevant in a commodity marketplace glutted with high-performance server choices?
What Should You Do?
Regardless of the effectiveness of these present IBM marketing strategies, if you're a current System i customer, you should take advantage of the opportunities that they afford.
This definitely includes PHP and MySQL. PHP is available now for System i, and MySQL will be available soon. Your skills and your knowledge of these open-source middleware services will help you educate your management on how they can potentially save money and reduce timelines for developing new applications.
It will also enable your IT organization to consider consolidating to the System i other applications that it may already be running under these services on Intel servers. It may even open a new paradigm for cross-platform development, in which IT builds applications on an Intel or Linux-based box and then consolidates those applications on the System i.
If all these promises hold—both IBM's and MySQL AB's—then MySQL could become YourSQL! And PHP could become your programming team's best-used tool. And your management will look down from on high, smiling at the System i, all the way to the bottom line.
Thomas M. Stockwell is Editor in Chief of MC Press Online, LP.