Security remains a big concern—tens of thousands of MongoDB servers were hit by a ransomware attack in early 2017. Exposing your database to the internet adds a vulnerability that many will find unacceptable. However, the latest breed of DBaaS systems has hardened interfaces that can mitigate most of these risks.
More significantly, the migration of applications into clouds may have created a tipping point for cloud databases. More new applications are being deployed inside the cloud than outside, creating a greater demand for cloud based databases.
Consequently, all database vendors are concentrating on DBaaS versions of their offerings. These include the usual suspects—Oracle, Microsoft, Datastax (Cassandra), and MongoDB. Amazon offers its own high-speed DBaaS options—DynamoDB (NoSQL), RedShift (SQL Data Warehousing), as well as hosted versions of Oracle, MySQL, SQL Server, and PostgreSQL. Microsoft offers the Cosmo DB-hosted DBMS as well as DBaaS versions of Microsoft SQL Server and various Hadoop-oriented data stores.
Google recently announced a publicly available version of Spanner—a distributed ACID-compliant SQL database. Spanner attempts to resolve the trade-off between availability and transactional integrity—the “CAP” theorem conflict—by coming down on the side of transactional integrity when forced, but doing everything possible to avoid any such compromise. Spanner employs atomic clock time synchronization to avoid some of the problems that occur in other distributed databases where machines on the network exhibit clock skew. Spanner also leverages the highly reliable Google network so that an availability outage, while not impossible, is at least unlikely.
Some of these cloud-based DBaaS systems—Cosmo DB, Spanner, and DynamoDB—now offer significant advantages over their on-premise counterparts. Network latency will continue to be an issue for an on-premise user or application, but with the bulk of applications now in the cloud, this is often not a concern. Security for these native DBaaS systems is strong, and they can offer throughput and scalability characteristics beyond what would be available on premise.
The biggest challenge facing Amazon, Google, and Microsoft is that customers are reluctant to accept the complete lock-in represented by these cloud database systems. An application based on these databases is very hard to migrate to another cloud—and, for this reason, only those completely committed to a specific cloud are likely to fully commit to a particular cloud vendor’s DBaaS.
Blockchain
Hype around cryptocurrencies and the underlying blockchain increased in 2017 to the point where regulators and governments have started to seriously examine the role of these technologies in finance and governance. Blockchains underlying Bitcoin, Ethereum, and other cryptocurrencies represent a sort of distributed, tamper-proof shared database with an immutable transaction history. While the blockchain itself cannot store large amount of data or support large-scale transaction processing, it has very significant implications for database management systems:
- Blockchains can be used to coordinate transactions across disparate databases, allowing for the first time a sort of reliable, two-phase commit that could allow a common version of the truth without complex intermediate and often manual processing.
- Blockchains show us a way in which data can be owned by individuals rather than by database administrators. One could use the blockchain to prevent any changes to personal information in a banking database without the customer’s express consent.
- Blockchains allow a way to ensure that database contents at any point in time can be proven to be consistent. Today, we rely on audit logs and complex security to ensure that the contents of the database have not been altered. The immutable blockchain ledger could be used to “seal” the contents of a database against such tampering.
What’s Ahead
The many compromises demanded by our current plethora of database technologies make selecting a database system far harder than it ought to be. Whichever DBMS you choose, you are probably going to experience a “not-quite-right” solution. Therefore, what we will see more than anything else over the next few years in the database marketplace is consolidation.
There are clear signs of this already: Virtually all database systems now support JSON in a native format and almost all systems—even those once called NoSQL—support the SQL query language. Columnar, in-memory, and graph technologies are now found across an increasingly broad range of database systems. As this convergence increases, we are likely to see more and more “multi-model” and “hybrid” database architectures. However, this convergence is probably going to play out over a very long period of time, and it’s hard to envisage an immediate future in which “one size fits all.”