One of the fiercest fights now occurring across the cloud is a battle for control over managed database services. The trouble is that rather than being a fight between traditional managed service providers, most of the conflict is between database vendors trying to sell managed database services directly.
For example, this week IBM announced it is making available a fully managed Db2 relational database service. Earlier in the week MongoDB announced it is extending its Atlas managed service for deploying instances of its namesake document database to Microsoft Azure and Google Cloud Platform (GCP).Atlas is currently hosted on Amazon Web Services.
Those offering now compete with managed database services from AWS, Oracle, Microsoft, Google, Redis, Snowflake Computing, and a small army of other providers. In each case, a vendor that makes a database has significantly expanded its managed service portfolio. In fact, Oracle reported this week that its cloud services revenue has topped $1 billion, which includes fully managed instances of the Oracle database and associated middleware technologies.
In general, there’s a battle royale happening across the entire database category, and variants of NoSQL databases are challenging traditional relational databases for supremacy. Of course, there’s really no such thing as NoSQL. The term itself describes a variety of database engines spanning everything from document and key value store databases to graph databases and instances of Hadoop. Confusion stems from the fact that despite the NoSQL moniker, all of these platforms support some form of SQL as the primary interface for interrogating data.
In theory, a key/value store database is different from a relational database in that it is optimized for massive amounts of unstructured data. In contrast, a relational database is optimized for structured data. Things get fuzzy, however, when you consider that most data is unstructured before it becomes structured, which means there’s a lot of data where the need for a relational database could simply be bypassed. Of course, relational databases have been employed for decades, so the number of applications that are optimized for a structured approach to managing data is in the millions. It’s also worth noting that the amount of data that can be stored in a relational database is increasing.
Database opportunities for MSPs
In practice, most organizations are going to wind up implementing a variety of databases, and therein lies the opportunity for MSPs. MSPs can basically make two choices. They can resell a managed database service provided by a vendor, or they can build their own. Competing head to head against a database vendor is a challenge. But, if most customers end up employing multiple database engines they are going to look for MSPs that have expertise across multiple platforms.
Databases are only a means to a larger data science end. If data is the new oil, then the database is the place where that oil gets refined. Entities that refine oil have historically made significantly more money than those that drill for it. In this context, MSPs need to decide if they simply want to own the data equivalent of a gas station or become something more akin to an oil-and-gas producer.