The Questions Nobody Writes Down: How Sivasankararao Kavuri Turned the Hidden Work of SAP Migration Into a Public Playbook

—

There is a kind of knowledge that lives only in the heads of the people who do the work. On an enterprise data migration, it sounds like this: “Which value mappings did we move to the test system last week?” “Why did this customer load with the wrong ship-to address?” “Who changed the cross-reference table?” These questions decide whether a go-live succeeds. Yet they almost never make it into a textbook, a white paper or a conference keynote.

Siva Kavuri has made a habit of writing them down.

Over the past three and a half years, Kavuri has published seven research articles on SAP data migration and integration. Read together, they amount to something unusual in this field: an honest, public record of where migrations actually break, and a set of methods for keeping them from breaking. With three new papers since October 2021, his work has shifted from the big architectural questions toward the details that experienced teams know matter most.

This article takes a different route through that body of work. Instead of a list, here are the questions his research sets out to answer.

“Why does every migration start so slowly?”

Kavuri’s first paper, published in April 2019, began at the beginning. RDS Content Installation Strategies: Accelerating S/4HANA Migration from Legacy ECC Landscapes looked at the weeks many teams lose simply getting SAP’s pre-built migration content running. His answer was to treat installation as a planned piece of engineering rather than a setup chore, so that teams reach their first trial load sooner and with fewer surprises.

“Why do the same defects keep coming back?”

A year later, in May 2020, Defect Lifecycle Management Across Mock, SIT, and UAT: An SLA-Driven Framework for Migration Quality took on a problem every project manager recognizes. A data error is fixed in one test cycle and quietly returns in the next. Kavuri proposed one lifecycle for every defect, with an owner, a severity and a resolution target, carried from the first mock load all the way to cutover. Quality stopped being a feeling and became a number.

“How do we connect the old world to the new one?”

In 2021, his attention turned to the cloud. Bridging On-Premises and Cloud: A Framework for HANA Smart Data Integration in Multi-Stream Data Lake Architectures (March 2021) offered a single, repeatable design for pulling several very different source systems into one cloud data lake. Four months later, Flowgraph-Driven Replication: Optimizing SAP HANA SDI for Real-Time Data Lake Ingestion on AWS (July 2021) showed how to make that pipeline run close to real time. Together, the two papers gave hybrid landscapes a blueprint instead of a collection of one-off scripts.

Those four papers built his reputation. The three that followed are what make his work stand out today.

“Migrations rarely fail on the big decisions. They fail on the details nobody wrote down.” 

Why does a load that worked in testing fail in production?”

Anyone who has run a large migration knows the sinking feeling. The data loaded perfectly in the development system and again in quality assurance. Then, in production, records fail or land with the wrong values. Very often the cause is not the data at all. It is the value mappings: the rules that translate a legacy code into its new equivalent, which drift apart as they are copied from one system tier to the next.

In October 2021, Kavuri published Cross-Environment Value Mapping: A Governance Model for Multi-Tier SAP Landscape Migrations. The paper treats value mappings as governed assets, not loose configuration. It lays out how mappings should be owned, versioned, approved and promoted from development to test to production, so that every environment runs on the same rules at the same time.

The idea sounds modest. Its effect is not. A single mismatched mapping can misroute orders, misstate balances or block thousands of records at cutover. By giving teams a governance model for something that is usually managed by memory and spreadsheets, Kavuri addressed one of the most common and least discussed causes of migration failure

“Is our landscape actually ready for the next load?”

His June 2022 paper goes after another piece of hidden plumbing. Cross-reference tables, known on projects as XRef tables, record how each legacy key maps to its new key in the target system. Every later load depends on them. A customer cannot receive an open order if the system cannot find which new customer number belongs to the old one.

XRef Table Redesign: A Systematic Approach to Data Migration Landscape Readiness argues that these tables are too often built ad hoc, refreshed by hand and trusted without checking. Kavuri sets out a systematic redesign: a consistent structure, a defined refresh process, and readiness checks run before each load cycle begins. The paper reframes “readiness” as something a team can verify, not simply assert, before it presses go.

It is the kind of work that rarely earns attention because, when it is done well, nothing goes wrong. That is exactly why it matters.

“How do we keep complex business relationships intact?”

The newest paper, published just last month, tackles one of the most intricate areas of SAP master data. In SAP Master Data Governance (MDG), customers and vendors are not simple records. A single customer may have several ship-to locations, preferred carriers, sales contacts and a place in a parent-and-child hierarchy. A vendor may invoice through a separate party. These links are expressed through partner functions, and getting them wrong breaks everything from shipping to tax to payment.

Partner Function Extension Patterns: Managing Complex Customer and Vendor Hierarchies in MDG (September 2022) catalogs repeatable patterns for extending partner functions across these hierarchies during migration. It shows how to load relationships in the right order, how to extend ship-to and carrier relationships to the right sold-to parties, and how to keep intercompany and invoicing-party links consistent. For teams moving to MDG, it turns a notoriously fiddly task into a set of known moves.

Why this body of work matters

Step back, and a pattern is clear. The early papers answered the questions executives ask: how fast, how reliable, how connected. The newer papers answer the questions practitioners ask late at night before a cutover: which mapping, which key, which relationship.

That shift is the heart of Kavuri’s contribution. Much of the writing on enterprise data focuses on platforms and strategy. Far less of it explains the governance of small things, even though small things are where projects fail. By putting value mappings, cross-reference tables and partner functions under the same research lens once reserved for architecture, his work raises the standard for what careful migration practice looks like.

There is also a generosity to it. Each paper takes knowledge that could easily stay private, the kind that makes one person indispensable on a project, and turns it into something any team can pick up and use. In a field where hard-won lessons are often lost when a project ends, that choice to publish is itself a contribution.

The timing could hardly be better. With thousands of companies working toward the end of mainstream ECC maintenance and more of them adopting MDG to govern their master data, the problems Kavuri writes about are landing on more desks every month.

Seven papers in, the through line is simple. Migrations rarely fail on the big decisions. They fail on the details nobody wrote down. Siva Kavuri has spent the past three and a half years writing them down.

THE SEVEN PAPERS AT A GLANCE

—

This content is brought to you by Marcy Betterly

Photo provided by Sivasankararao Kavuri