About UDDI root registry and affiliate Registry

The V3 specification proposes a new UDDI system framework. the V3 specification assumes that there will be many different UDDI registries, and each UDDI registry is composed of many or one node (which can be seen as equivalent to the operator site in the V2 Specification. these registries form a large system. if the V2 standard system framework is still used, it is difficult to ensure that the key-value keys of all registries do not conflict with collision. the solution is to create a UDDI root registry that manages other registries and these registries become the root Registry's affiliate (affiliated) registry. for example, a root registry is a UDDI business registry (UBR), which has a mechanism to generate a unique uuidkey and verify the validity of the domainkey through a digital signature. these policies guarantee the global uniqueness of domainkey and uuidkey in the whole UDDI business registry, and this UBR is theoretically a root registry.

Shows the UDDI architecture framework under the V3 specification:

The architecture of V3 enables UDDI to have many functions that were not implemented by the previous architecture, such as a publisher

You want to copy all the UDDI registry entity registered by one registry to another registry while maintaining the same key value. in addition, V3 also provides data sharing for wider distribution. v3 specifications are very important for the development of UDDI.

The key concept is described in detail here.

In the V3 version, because of the concept of multiple registry centers, we need to define a key that can be unique in all registries so that our data can be freely moved. The format of this key: UDDI: partition (1. N): Key. How is this key mechanism implemented:

We have a root UDDI business registry, which works together with IBM, Microsoft, NTT communications, and sap. In this way, other registries are assigned a value by the root Registry. In the specification, this value refers to a partition (domain ). That is to say, this root registry will assign a value to different sub-registries. This sub-Registry can only publish its own key in this value or generate its own next level of partition.

For example, for IBM, the root registry pays it a value: COM. IBM. This is the part of IBM. In this way, in this sub-registry, IBM programmers do not need to care about the registry information of other companies, that is, they will think that the COM. IBM registry is their root. In this way, the administrator of the registry can create the partition of his company in the Registry and assign it to different service publishers. The key can also be directly published in this domain. For example, we can publish a service to complete the login function of the IBM website and set key = login for it. At this time, the complete value of this key should be the value added to its domain, separated by a colon in the middle: UDDI: COM. IBM: login.

For example, the administrator of the sub-Registry com. IBM has created a new partition and paid it to the Development Department: develop. At this time, the publisher of our development department can create a key in this domain. To log on to the development department's own website, we can publish a login service, whose key is also called login. The complete value of this key is UDDI: COM. IBM: Develop: login. In this way, in the Development Department domain, the key of the login service is login. Its value is unique in the Development Department domain. In the IBM domain, the Development Department's service, we add the develop domain name before this key value, so that our key becomes the develop: Login, in this way, we have ensured that this key is unique within IBM. similarly, in the root registry of the whole world, the key of our service becomes COM. IBM: Develop: Login ensures that our keys are not repeated anywhere.

At the same time, the publisher can create its own sub-partition, and create a sub-partition on the sub-partition, so that infinite recursion can be performed. Now we can create a tree-based partition structure. We can release keys on any node of this structure, so they will not be repeated.