Recently, Microsoft has announced a product that seems to blend together to give end-to-end analytics at a very large scale on its cloud platform with Synapse Analytics.
Over the years, we have seen mass migration to the cloud as businesses signed up by the thousands to capture the advantages of flexible, large–scale computing and data storage. But in the last five years or so, what we saw was that the next iteration of that tech revolution, in which companies would use their growing stores of data to get more tangible business insights that had slowed down.
Data is spread across big data lakes, data warehousing, data pipelines and it’s difficult to translate the benefits of data into actionable analytics. Recently, Microsoft has announced a product that seems to blend all of that together to give end-to-end analytics at a very large scale on its cloud platform with Synapse Analytics, which is a rebrand of Azure SQL Data Warehouse — first announced in 2015.
What’s Unique About Azure Synapse Analytics?
Synapse Analytics aims to unify a range of analytics workloads, including data warehouse, data lake, machine learning and data pipelines that act as a bridge between all those things. The core data warehouse engine has been ramped up with novel features to go against other cloud data warehouse platforms directly, including the ability to include workloads through explicitly provisioned or on-demand (serverless) infrastructure, each with its specific pricing model.
The biggest highlight is the integration of Apache Spark, Azure Data Lake Storage and Azure Data Factory with a unified web user interface. The powerful combination of Spark with Azure Data Lake Storage (ADLS) and Azure Data Factory together on the UI, gives users the control over both data warehouse/data lakes and accommodate data preparation and management.
Another great thing about Azure Synapse is that organisations that are using it can have project timelines measured in hours and not weeks or months. That’s due to the fact that Synapse uses Transact-SQL (T-SQL) language for querying both relational and nonrelational data at petabyte scale. In an on-stage demonstration, Synapse took 9 seconds to perform a complex query on a petabyte-scale dataset. When the same query on Google Big Query, it took a long time, around 11 minutes, according to Microsoft. In this particular incident, the query on Synapse was 75 times quicker than when it was performed on Google’s BigQuery.
Microsoft’s Synapse Compared With AWS’ RedShift & Google’s BigQuery?
Azure Synapse also integrates with Power BI and Azure Machine Learning to gain insights for all users, all the way from data scientists to the business users using Power BI. Microsoft also said Synapse partner ecosystem that includes Databricks, Informatica, Accenture, Panoply, Talend, Attunity, Pragmatic Works and Adatis, many of whom hybrid cloud analytics providers who have a foot in the same market. It becomes clear that Azure Synapse is, therefore, a competitive product targeted at large vendors like AWS RedShift and GCP’s BigQuery.
Another big aspect apart from scale where Synapse Analytics takes the lead is concurrence. Microsoft demonstrated during Ignite that with as few as 50 users Synapse’s query response time was three times quicker than Amazon’s Red Shift and five times quicker than Google’s BigQuery. As it was scaled from 50 to 150 concurrent users, Red Shift started queuing requests at around 100 while BigQuery began failing requests above 100 users. Synapse, on the other hand, kept delivering without faltering, which is quite impressive.
The objective of analytics platforms is to deliver actionable insights and therefore what is unique about something like Synapse is the integration with Microsoft Power BI — a tool which is a part of the Office 365 Enterprise E5 version, catering to more than a billion enterprise employees worldwide. As a result, businesses can build Power BI reports and enterprise-grade semantic models right from within the Synapse workspace, which is where it can trump other players.
Cutting edge analytics platforms require predictive capabilities as well, which is why Microsoft decided to integrate Synapse with Azure machine learning, allowing users to easily scale out, build and operationalise complex machine learning models, right from the Synapse workspace. During a demo, Microsoft modelled a simple query that included native scale-out machine learning-based scoring on the Synapse nodes.
The cloud wars are intensifying on things like big data analytics and BI vendors, which means large vendors are upping their game constantly with new features. It also proves that despite huge acquisitions sprees like Tableau from Salesforce and Looker from Google may not be enough to compete with utility and ease of use as in the case of Synapse. Microsoft proved with the launch that innovative resources with existing products can satisfy customer needs, operationalise analytics more fluidly, and give security and compliance to enterprise users more smoothly.
Enjoyed this story? Join our Telegram group. And be part of an engaging community.