Dependencies are easy to underestimate until one breaks. Most enterprise teams can name the main applications that use their ASE servers, yet far fewer can list the scheduled exports, cross-server links, and forgotten integrations quietly relying on them. Sybase database consulting helps close that gap, turning a vague sense of “what probably connects to what” into a clear, reliable record.
To show how that works in practice, this article follows a composite example drawn from situations common across ASE environments. The organization is illustrative, but the problems and the methods used to solve them are very real.
A Familiar Starting Point
Picture a mid-sized insurer running three production ASE servers. One handles policies, one handles claims, and one supports finance and reporting. The systems have grown over two decades. The original DBAs have moved on, and the documentation describes the environment as it looked ten years ago.
Leadership approves a plan to move the claims server onto new hardware. On paper, it’s a routine project: dump the databases, load them on the new machine, update the connection settings, and switch over on a Saturday night.
The Server Move That Exposed the Problem
The move itself goes smoothly. By Monday, though, three issues appear. The finance team’s overnight reconciliation report is empty. A partner portal that checks claim status is returning errors. And a reporting server that pulls claims data through proxy tables has stopped refreshing.
None of these connections were in the project plan. Each was added years earlier, for a sensible reason, by someone who no longer works there. This is the point at which many organizations bring in experienced Sybase database consulting, not just to fix the immediate breakages but to make sure the next change doesn’t repeat them.
What a Proper Discovery Turned Up
The consultants start by treating the environment as unknown territory and gathering evidence from several directions.
Inside the database, they review what ASE has recorded. The sp_depends procedure reports relationships stored in the sysdepends table, but that record is often incomplete, especially where objects were created out of order or code builds SQL dynamically. So they also scan the source text of procedures, views, and triggers stored in syscomments, looking for references to other databases and servers.
Next comes runtime evidence. ASE’s monitoring tables show which programs connect, from which hosts, and what they run. Collected over a full month, including month-end, this activity reveals connections that only appear occasionally, such as a quarterly regulatory extract.
Outside the database, the team checks interface files, cron schedules, replication definitions, and file transfer jobs using bcp. Then they sit down with application owners. Two of the most important findings come from those conversations rather than from any log.
Dependencies Discovered by Type
By the end of the discovery, the picture looks very different from the original documentation.
| Dependency type | What was found | Why it was missed |
| Cross-database references | Claims procedures reading policy tables directly | Not recorded in sysdepends |
| Cross-server links | Reporting server using proxy tables to claims | Configured on a different server |
| Scheduled jobs | Finance export run by a cron job on an old host | Lived outside the database entirely |
| Application connections | Partner portal with a hard-coded server name | Built by an outside vendor |
| Replication flows | A warm standby fed by Replication Server | Documented, but not linked to the move plan |
| Shared server objects | Logins and user-defined data types the claims databases relied on | Defined at server level, not in application docs |
The table makes one lesson clear. The riskiest dependencies weren’t the complex ones. They were the ones sitting outside the place anyone thought to look.
Building a Map the Business Could Use
Finding dependencies is only half the job. The consultants organize the results into a dependency map that records each connection, its direction, how often it runs, who owns it, and what would break if it stopped.
They also rank every link by business impact. Billing and regulatory feeds go to the top. A test script that hasn’t run in two years is flagged for removal. In this example, a noticeable share of the connections turn out to be unused and are retired, which simplifies the environment immediately.
To keep the map from going stale, it’s tied to change management. Any new proxy table, scheduled job, or application connection must be added before it goes live, and each entry has a named owner. A short quarterly review compares the map against fresh monitoring data to catch anything that slips through.
How the Map Changed Later Projects
The real return shows up over the following year. A planned ASE patch is tested against exactly the right applications, with every owner notified in advance. When the finance server is consolidated with the policy server, cross-server links are rebuilt before the move rather than repaired after it.
Most significantly, the insurer begins planning a gradual migration of its reporting workloads to SQL Server. Because every connection is already mapped, the migration plan starts with a complete list of what must be rebuilt, redirected, or retired, instead of weeks of rediscovery.
Lessons That Apply Beyond the Example
The details will differ from one organization to the next, but the pattern repeats almost everywhere. Documentation tends to describe how systems were designed, while dependencies reflect how they were actually used. The gap between the two grows a little with every quick fix, new report, and staff change.
Two habits make the biggest difference. The first is gathering evidence from running systems rather than relying on memory or old diagrams. The second is giving every dependency an owner, so there’s always someone to ask before a change goes ahead. Organizations that adopt both find that upgrades, moves, and migrations stop feeling like gambles and start feeling like ordinary planned work.
Frequently Asked Questions
Is sp_depends enough to find dependencies in ASE?
It’s a helpful start, but it often misses references. Source code scans and runtime monitoring are needed to build a complete picture.
How long should monitoring run during discovery?
At least one full business cycle, including period-end processing, so connections that run only occasionally are captured.
Should dependency mapping happen before a migration?
Yes. A complete map is one of the most valuable inputs to any migration plan, since it lists every connection that must be handled.
The Last Word
The lesson from examples like this is consistent: dependencies rarely cause trouble because they exist, but because nobody can see them. Sybase database consulting brings them into view through code analysis, runtime monitoring, and conversations with the people who depend on them, then helps keep that knowledge current as systems change. For organizations planning upgrades, consolidation, or a move to SQL Server, the team at ralantech.com offers ASE consulting and migration experience that can help turn a tangled web of connections into a usable plan. A practical first step is collecting a month of connection data from your busiest ASE servers, since it almost always reveals links nobody expected.
