In enterprise space we talk a lot about the "system of records". You cannot complete a brainstorming session without mentioning it in some context. SAP is a system of records for material master and product BOM, Seibel has customer info, Ariba has supplier inventory and so it goes.
There is a similar honorary status at the infrastructure level which is usually given to the most popular component. Be it Apache for web server, Tomcat for servlet container etc. They constitute system level services for any enterprise stack. They are fairly standardized to the extent that they are taken as obvious choice during architectural design.
Some where between system services and business system of records lies another set of service – in some cases I would say cluster of services. LDAP is the oldest member of this service category. It acts as a system of records for access and identity information. Microsoft Active Directory, Novell, Sun and Netscape LDAP lead the pack in this category. There are other services in this category which include – single sign-on, federated process orchestration, portal services etc.
If you follow the web services growth then you will realize that the identity management is key to a successful web services rollout. Identity and access management are closely related – they are like yin and yang. They have to co-exist. They do this on top of LDAP persistence layer. That alone is a reason to take LDAP as a critical component of SOA-enabled enterprise architecture.
SOA infrastructure build-out is what makes Redhat’s decision to acquire Netscape LDAP code (and subsequent decision to open source it) a very smart move. By tightly integrating it with Redhat Linux Enterprise server this will provide a very compelling foundation level architecture for open source based SOA.
The fact that Netscape LDAP is of high quality and can claim most intense deployments out there will help position Linux favorably. Watchout Sun. As in other stories related to Sun, irony is that Sun LDAP server and Netscape Directory server share some common code. Now this same code will be available under CDDL and GPL separately.
Lets see how Redhat $25million gamble changes the enterprise architect’s mindset in deciding which stack provides the biggest bang for the buck.