Visit most any Fortune 2000 company these days that spends a lot of money on relational database software and chances are you’ll find a proof-of-concept project in which the internal IT organization is, at the very least, experimenting with some form of a so-called NoSQL database to run an online transaction processing (OLTP) application.

Once the exclusive preserve of relational databases from Oracle (ORCL) and IBM (IBM), NoSQL databases such as the open source Apache Cassandra project are starting to gain some traction. For example, DataStax, which provides commercial services around Cassandra, now claims to have more than 1,000 customers. In fact, that momentum is one of the reasons managed service providers such as Datapipe have created a managed services practice around Cassandra.

It’s still too early to say how much NoSQL databases such as Cassandra will cut into sales of traditional relational databases. In fact, in addition to experimenting with Cassandra many of those same IT organizations at the high end of the market are also trying to figure out what percentage of their application workloads they might want to move into an in-memory database.

For the most part, NoSQL databases come in two forms. There are document databases aimed mostly at workflow applications. At the high end of the market, however, there are NoSQL databases specifically optimized for transaction-processing applications that are starting to make their presence felt in the enterprise.

Matt Pfeil, chief customer officer for DataStax, noted that a big part of the appeal of Cassandra—besides the open source nature of the project—is that Cassandra is designed to be able to run transactions at scale using commodity x86 hardware that IT organization can scale out as needed. In contrast, most of the existing relational database deployments are based on much more expensive servers.

Naturally, the deployment of relational databases on those servers historically has generated massive amounts of profits and revenue for solution providers. If enterprise customers decide to shift application workloads to Cassandra in large numbers, the impact of those decisions will go well beyond the number of database licenses being sold by Oracle and IBM. Solution providers with relational database and server expertise might experience a sudden drop-off in demand for their services as well.

Of course, none of this is going to occur overnight. But there is now enough buzz circling Cassandra for solution providers to not only take notice, but perhaps also start thinking about how to build their own Cassandra practice. The real irony here is that Cassandra is named after the daughter of King Priam of Troy, who foretold of the coming doom of the city, but was cursed by the gods to make sure that no one actually listened to her. While most solution providers may not be familiar with Cassandra, they do at very least know what ultimately became of Troy.