What is Transaction Path Analysis?

Question

I came across the term Transaction Path Analysis which I am not familiar with. A Google search resulted in only a few results, most of which seem to assume everyone knows what it means. From these my understanding is that it is about tracking what columns are used by queries to access data, in preparation for deciding indexes. This seems like common sense, but I'm surprised I can't find more specifics on it.

Is there a concise definition for transaction path analysis, a process for completing one, or any alternative more commonly used phrases that mean the same thing?

Answer 1

This term relates to the process of identifying the significant events in the transaction life-cycle, including error-processing, auditing, save-points, nested transactions, and distribution paths for distributed transactions.

I think you may find more results with the synonymous term Transaction Life-cycle Analysis.

Answer 2

Perkinson's mtehodology merges logical(ideal) process model and logical(ideal) data model into one combined ideal model and calculates resources of each access path in a table. In this calculation, peak event frequency, entry point, data volume, cardinality, from entity, to entity, access type & required resources by each access, etc. are needed. According to calculation result, we can add new access path or new entity type for optimal performance. I strongly recommend you to read J. Palmer & S. McMenamin's book 'Essential Systems Analysis' for the exact meaning of 'Logical'.

Answer 3

Richard's goal in his methodology is to arrive at four characteristics of the system under study to correctly design the logical and physical database. They are (in sequence of importance to the methodology)

1. Entry,
2. Structure,
3. Load and
4. Direction.

All of these are identified and more importantly are quantified in the Transaction Path Analysis stages of his method, except Structure.

Before Dick, most methodologies stopped at identifying the overall underlying logical Structure of the data, the entities, their keys and their relationships, and from that they designed the database. Dick identifies these in the first two, of the four total, stages in his method. But he would say, "it's not enough to know just what the data is, and how it relates. We also need to know it's usage."

In his four stages you learn, in sequence,

1. What is the data?,
2. How does it relate?,
3. How us it used?, and
4. How much?

The third and fourth stages are Transaction path analysis Qualitative, Stage 3, and quantitative, Stage 4. I've had the very good fortune to have worked For and With Dick for five years at QED Information systems in Wellesley MA. I often told my students, "I'm not one of the greats, but I AM a 'ventriloquist dummy' for one of the greatest, Richard Perkinson." Any of my success in the IT industry can be attributed to his teaching, guidance, and faith in me.