Identification and Match Criteria Configuration
Depending on your data source, available credentials, and specific system security settings, an integration point may have access to only a limited set of attributes when identifying a CI.
For example, IP range discovery detects two IP addresses (10.12.123.101 and 16.45.77.145), and creates two nodes. However, detailed system discovery may detect that those two IP addresses are actually configured on two network interfaces in the same node.
This means that you cannot always rely on a single matching set of attributes for identification – other possible attributes that can potentially help to identify the CI should also be listed. In the previous example, the node identification attributes can be the IP address and the network interface. If you use the IP address to identify the CI, you see that all three discovered nodes are the same node.
Suppose that detailed system discovery detects a node with IP address 10.12.123.101 and network interface MAC1. At some point, that node was shut down, and the same IP address (10.12.123.101) was given to another node with network interface MAC2. These two nodes have the same IP address; however; it is obviously not the same CI. Performing match validation on the network interface data helps us to realize that it is not the same node.
The identification criteria are used to select candidates, and the match criteria are used to approve the identification result or dismiss it. For example, while handling input CI A, we may get identification candidates B and C, and the match criteria will dismiss B. In that case, we are left with C, which means that A is identified as C.
Identification Criteria
Data that the reconciliation engine receives from different data sources may contain different subsets of the attributes (topology) necessary for identifying a CI. The identification criteria should contain all potential attributes on which CI matching can be done.
Each identification criterion defines a potential condition for CI matching. The criterion can be an attribute such as node name, or topology such as IP address. A criterion may contain two or more conditions, to create a more complex matching rule. It may also contain different condition operators such as equals or contains, or it may contain some master value that defines a value in the CI that will always allow a match.
During the identification process, all identification criteria are running to find all candidate CIs for matching.
Possible Node Identification Criteria
-
HW ID
-
Network interface (containing a condition operator)
-
Node name
-
IP address (containing a condition operator)
Match Criteria
While identification criteria list all potential attributes for matching the data, match criteria contain the attributes that are essential for matching CIs, if any exist. This means that if two CIs are marked as candidates to be matched by the identification criteria, the match criteria will check if the data exists in both CIs in order to match the condition.
Match criteria are also used during the Data-In process in case of multiple matches, to make the decision to merge CIs from the CMDB. The CIs are merged only if the match criteria are satisfied. If one of the CIs does not satisfy the match criteria, the merge is not performed.
Specifications
A match criterion is satisfied if two candidate CIs have the same essential data (as defined in the that criterion), the data matches the condition, or if at least one of the CIs has no essential data.
Match criteria can be divided into two categories:
- Match verification criteria – if the verification criterion is not satisfied on two candidate CIs, these CIs are not matched.
- Match validation criterion – if the criterion with higher priority is satisfied (without missing data) on two candidate CIs, the validation criterion with lower priority is even not checked and the CIs are marked as matched. Similarly, if the validation criterion with higher priority is refuted on two candidate CIs, the criterion with lower priority is even not checked and the CIs are marked as not matched.
Possible Node Match Criteria
- Match verification criteria uses the discovered OS data for verification. This means that if two nodes have discovered OS data and this data does not match, these two nodes are not matched.
-
Match validation criteria (ordered from highest to lowest priority):
Priority 1. HW ID with an equals operator
Priority 2. Node name with an equals operator
Priority 3. Network interface with a contains operator
This means that if two nodes with the same HW ID are discovered, they are marked as matched even if they have different network interfaces or node names. On the other hand, if the discovered HW IDs on the nodes are not the same, the nodes are not marked as matched even if the network interfaces and node names are the same. The network interface rule is checked only if one of the nodes has no discovered HW ID.
