To further strengthen our commitment to providing industry-leading coverage of data technology, VentureBeat is excited to welcome Andrew Brust and Tony Baer as regular contributors. Watch for their articles in the Data Pipeline.
At the PASS Data Community Summit in Seattle today, Microsoft announced the release to general availability (GA) of SQL Server 2022, the latest version of Microsoft’s flagship relational database platform. The private preview of SQL Server 2022 was announced about a year ago; the first release of SQL Server on Microsoft’s Windows Server operating system was released some 30 years ago. That platform’s seniority notwithstanding, it has been continuously modernized and the 2022 release of the product is no exception.
VentureBeat spoke with Rohan Kumar, Microsoft’s Corporate Vice President, Azure Data, for the business perspective, and Asad Khan, Vice President, Azure Data Engineering, for the technological details. Kumar announced the GA news during his keynote address at the PASS event this morning and spoke to VentureBeat about the business value of the release; Khan talked the technology details.
Hybrid cloud or versatile cloud?
On the business side, Microsoft sees this release of SQL Server as the one that leverages numerous components of Microsoft’s Intelligent Data Platform (IDP) and, despite it being an on-premises product, the most cloud-oriented version of SQL Server yet released. Both of these pillars are undergirded by integration with Azure Synapse Analytics, Azure SQL Database Managed Instance (MI), Azure Active Directory and Microsoft Purview. The cloud story is further substantiated by compatibility with S3-compatible object storage, although that has significance on-premises as well.
At a high level, though, Microsoft is looking to SQL 2022 as the release that brings cloud innovations back to the customers who still need to run on-premises. Its high degree of compatibility with cloud releases of SQL Server, which includes numerous editions of Azure SQL Database — especially Azure SQL Database MI — means that on-prem customers can gain access to the features of the cloud releases. It also makes it easier for customers to move to the cloud when they’re ready to do so and it allows Microsoft to work with those customers and understand better the factors impeding their move to the cloud.
Intelligent Security Summit
Learn the critical role of AI & ML in cybersecurity and industry specific case studies on December 8. Register for your free pass today.
The GA of SQL 2022 also parallels the maturation of the operational-analytical data technology balance. Kumar explained that the days of throwing operational databases, business intelligence, analytics and data governance at customers as rather separate components, are over. Instead, Microsoft is now working to stitch all these things together, and robustly support hybrid cloud scenarios as it does so.
Microsoft has worked hard to earn certifications and compliance under various government and industry regulatory frameworks. By doing that and making the on-premises and cloud products more compatible and interchangeable, Microsoft wants to remove most — or even all — of the friction in moving workloads to the cloud. It’s even making it easy to move them back on premises, if that ends up being a priority. Knowing that a move to the cloud isn’t irrevocable may just make customers more confident moving most of their workloads there.
This isn’t an Azure-specific premise, either. During our briefing, I asked Khan if disaster recovery scenarios could work not only between SQL Server and Azure SQL Database MI (as I’ll detail shortly), but also with Amazon RDS, once SQL Server 2022 is available on that platform. Not only did Khan say it would, but he said that kind of thing is the very point of the release, and not just some curious edge case.
Yes, many enterprise organizations run workloads in a hybrid mix of on-premises and cloud platforms, but maybe what they really want is for the infrastructure to be fungible and interchangeable, so the workloads can go anywhere, and be moved anywhere else. That ideal seems to underly the strategic direction for SQL Server 2022, at least rhetorically. As it turns out, the substance of the release supports that strategy too.
So what are the technical enhancements that underly these talking points and the strategic direction? To begin with, a new Link feature for Managed Instance means that SQL Server on-premises can now pair and serve almost interchangeably with Azure SQL MI. By using a simple wizard, database administrators (and probably non-DBAs too) can configure a provisioned MI in the cloud to serve as a secondary, failover node to a SQL Server 2022 instance on-premises, or vice versa. Furthermore, the provisioned MI can be used as a readable replica to distribute workloads, in addition to its fault tolerance role.
Next, using a feature called Azure Synapse Link, operational data in SQL 2022 can be replicated, silently and in the background, to dedicated pools (data warehouse instances) in Azure Synapse Analytics. The transaction log/change feed-based replication can happen on a continuous or scheduled basis. This feature was already available for Azure SQL Database (the mainstream cloud version of SQL Server) and, as of this writing, is still in preview on the Synapse Analytics side. It provides one of many options for SQL Server customers to pursue operational database and analytics workloads together.
Another such option is the enhancement of SQL Server’s PolyBase, a data virtualization and big data connectivity feature, to be compatible with Amazon S3 and all object storage systems that are API-compatible with it. Here again, the cloud/on-premises paradox raises its head as many S3 API-compatible storage platforms, such as Minio, run on-premises. As a result, Microsoft touts the new PolyBase as providing access to any data lake. This technology enables database backup to S3-compatible object storage, too.
Ironically, though, PolyBase will no longer support connectivity to on-premises Hadoop clusters. But, as a result, PolyBase’s dependency on the Java runtime has been eliminated, which raises the prospect that more customers may install it. If so, it would probably be a good thing for SQL Server’s integration with the modern open source data analytics stack, much of which is in the cloud.
So Synapse Link provides export connectivity to data warehouses and PolyBase provides import connectivity for data lakes (and export too, via the new CREATE EXTERNAL TABLE AS SELECT — CETAS — command). But what if customers want to perform analytics on SQL Server itself? There are new capabilities here as well, in the form of enhancements to columnstore indexes, which are designed for operational analytics. The short version of the enhancement is that it makes operational analytics faster. The longer version is that a new capability allows clustered columnstore indexes to be physically ordered, enabling something called “segment elimination.” Segment elimination lets SQL Server skip over whole batches of data that are not relevant to a query, rather than having to scan all that data and determine its irrelevancy by brute force.
SQL Server 2022 also includes enhanced support for JSON data, query intelligence for better performance and a Ledger feature that enables blockchain-style tamper-proofing in the database. There’s also integration with Azure Active Directory for authentication, Microsoft Defender for security and Microsoft Purview for access permissions, data classification, data cataloging, and data lineage.
I’ve used SQL Server since the early 1990s. This new 2022 version adds support for important new cloud, database and analytics technologies while it maintains consistency with, and fidelity to, the classic platform that has a huge community of skilled professionals. While Microsoft pursues newer platforms like its Azure Cosmos DB NoSQL platform and supports open source databases like PostgreSQL, it never seems to lose faith, or abandon investment, in SQL Server. The market seems to reward Microsoft for this policy. It will be interesting to see what SQL Server’s 4th decade may bring.