DATA IS THE backbone of business—and managing and analyzing it has become more critical than ever.
Take, for example, Brazil’s Itaú Unibanco, the largest bank in Latin America. In 2019, the company upgraded to the latest version of Microsoft’s SQL Server running on Intel technology. The bank oversees sensitive information and handles roughly five billion transactions per day. Doing this requires massive amounts of processing power, and changes to the data need to be tracked and remain secure. Managing this level of information can overwhelm even advanced systems, so Itaú Unibanco upgraded to SQL Server 2019 with the newest Intel Xeon Scalable processor–based servers.
Then something shocking happened.
The bank noticed a massive increase in performance, up to 80 percent, with a 20 percent reduction in processing time, thanks to their Intel-based SQL Server platforms. Even more amazing, these improvements were achieved without having to change the underlying code—something that’s often required to make the most of a new system. The IT team reached out to Microsoft with one question.
What’s going on?
“They asked, ‘What the heck did you guys do?’” laughs Bob Ward, Principal Architect for the Microsoft Azure Data team, which oversees the SQL Server and Azure Data technology architecture. “It’s a testament to our engineers, who continually enhance the database engine with built-in query intelligence so that customers can gain performance with no application changes required.”
Microsoft is hoping to bolster this performance advantage with the newly announced SQL Server 2022, which—with the help of Microsoft and Intel engineers—will be optimized for the Intel 3rd Generation Xeon Scalable processors. The system, available in 2022, is part of Microsoft’s broader vision for an end-to-end data platform providing customers with operational databases, analytics and governance, and the ability to seamlessly bridge on-premises deployments to the cloud. Overall, it includes even more game-changing upgrades to query processing, security, and disaster recovery—providing a holistic opportunity for companies to harness data and leverage it for growth.ADVERTISEMENT
“SQL Server 2022 will be revolutionary,” says Pedro Lopes, Principal Architect for the Microsoft Azure Data team. “It’s taking us and our clients leaps and bounds ahead into the future.”
A Data Platform that Keeps Up with a Changing World
Microsoft has been a leader in data management for more than 30 years, and the SQL Server engine has evolved with its customers to meet them where they are, offering solutions on-premises, in the cloud, at the edge, and in hybrid environments. SQL Server 2022 builds on that foundation. This latest release is more connected to the cloud, through Azure, than ever. This includes disaster-recovery capabilities with connections to Azure SQL Managed Instance, analytics processing through Azure Synapse Link, and data-governance policies through integration with Azure Purview. Paired with Azure, SQL Server 2022 works seamlessly in connection with the cloud.https://view.ceros.com/cn-live/wired-microsoft-intel-2?heightOverride=1009
“Our driving force is to ensure a customer’s data is where they need it to be, whether that’s due to customer preferences, regulatory requirements, or just ease of use,” says Lopes. “With every version of SQL Server, we want to give them the tools to gather and explore the data wherever they need it, all while keeping it safe.”
To optimize SQL Server 2022 on a hardware level, Microsoft engineers worked closely with Intel. As a result, one innovative outcome is the use of Intel® QuickAssist Technology (QAT) for the SQL engine, a feature that offloads data compression on SQL Server bare-metal platforms or virtual machines. This data-compression technology enables faster online transactional processing (OLTP) operations in SQL Server while reducing the impact on the system’s transaction throughput. Users can also expect database backup speeds that are 230 percent faster than with previous versions, which is critical to helping database administrators and developers get to production quicker.
SQL Server 2022 also enhances digital security—a paramount need for businesses today. Global losses from cybercrime totaled more than $1 trillion in 2020, according to a report by the Center for Strategic and International Studies, double the amount from only two years before. Over the past ten years, SQL Server has had fewer vulnerabilities than the competition1. To continue keeping data secure both in the cloud and on-premises, SQL Server 2022 includes several innovations. One is a new feature called SQL Ledger. Ledger uses the concept of blockchains built into SQL Server to track changes to the database using crypto hashing. This feature tracks any potential data changes from malicious actors while also enabling third parties to confirm if anything has been compromised. In regulated industries like finance and insurance, where auditors need to verify a database’s integrity, this is a vital feature.
“The great thing is that you can always backtrack to a previous version of the data, because it will have been moved to an immutable ledger,” says Lopes. “This is a very practical use of blockchain technology that will have huge benefits for companies.”
Without the ability to analyze data, though, information goes to waste. So SQL Server 2022 is introducing the next generation of Intelligent Query Processing (IQP) capabilities. One groundbreaking new feature is called approximate percentile. Organizations are inundated with torrents of data, which in turn create massive data sets that take an increasing amount of time and resources to analyze, even with advanced algorithms and processors. Previously, SQL Server 2019 helped mitigate this resource-usage issue by building in an “approximate count distinct” function, which quickly calculated approximations using a smaller memory footprint, rather than relying on exact operation equivalents, which can take significantly longer. SQL Server 2022 extends that by offering an additional approximate percentile equivalence, which allows for even lighter-weight calculations that fall within a reasonable error range. The feature is particularly useful for discovering trend lines in data, when it’s not necessary to crunch every single number before making decisions based on general trajectory.ADVERTISEMENT
“We’re continuing to expand the approximate QP space, because as data volumes get bigger, there’s still very little tolerance for latency,” says Joe Sack, Principal Group Program Manager at Microsoft. “Customers want their data quickly, especially the front office, and we can now get you within two percent of the exact value in a much shorter time frame to make decisions.”