Apache Spark is a multi-language engine for executing data engineering, data science, and machine learning on single-node machines or clusters.
N/A
Azure Synapse Analytics
Score 7.6 out of 10
N/A
Azure Synapse Analytics is described as the former Azure SQL Data Warehouse, evolved, and as a limitless analytics service that brings together enterprise data warehousing and Big Data analytics. It gives users the freedom to query data using either serverless or provisioned resources, at scale. Azure Synapse brings these two worlds together with a unified experience to ingest, prepare, manage, and serve data for immediate BI and machine learning needs.
Well suited: To most of the local run of datasets and non-prod systems - scalability is not a problem at all. Including data from multiple types of data sources is an added advantage. MLlib is a decently nice built-in library that can be used for most of the ML tasks. Less appropriate: We had to work on a RecSys where the music dataset that we used was around 300+Gb in size. We faced memory-based issues. Few times we also got memory errors. Also the MLlib library does not have support for advanced analytics and deep-learning frameworks support. Understanding the internals of the working of Apache Spark for beginners is highly not possible.
It's well suited for large, fastly growing, and frequently changing data warehouses (e.g., in startups). It's also suited for companies that want a single, relatively easy-to-use, centralized cloud service for all their data needs. Larger, more structured organizations could still benefit from this service by using Synapse Dedicated SQL Pools, knowing that costs will be much higher than other solutions. I think this product is not suited for smaller, simpler workloads (where an Azure SQL Database and a Data Factory could be enough) or very large scenarios, where it may be better to build custom infrastructure.
Quick to return data. Queries in a SQL data warehouse architecture tend to return data much more quickly than a OLTP setup. Especially with columnar indexes.
Ability to manage extremely large SQL tables. Our databases contain billions of records. This would be unwieldy without a proper SQL datawarehouse
Backup and replication. Because we're already using SQL, moving the data to a datawarehouse makes it easier to manage as our users are already familiar with SQL.
With Azure, it's always the same issue, too many moving parts doing similar things with no specialisation. ADF, Fabric Data Factory and Synapse pipeline serve the same purpose. Same goes for Fabric Warehouse and Synapse SQL pools.
Could do better with serverless workloads considering the competition from databricks and its own fabric warehouse
Synapse pipelines is a replica of Azure Data Factory with no tight integration with Synapse and to a surprise, with missing features from ADF. Integration of warehouse can be improved with in environment ETl tools
If the team looking to use Apache Spark is not used to debug and tweak settings for jobs to ensure maximum optimizations, it can be frustrating. However, the documentation and the support of the community on the internet can help resolve most issues. Moreover, it is highly configurable and it integrates with different tools (eg: it can be used by dbt core), which increase the scenarios where it can be used
The data warehouse portion is very much like old style on-prem SQL server, so most SQL skills one has mastered carry over easily. Azure Data Factory has an easy drag and drop system which allows quick building of pipelines with minimal coding. The Spark portion is the only really complex portion, but if there's an in-house python expert, then the Spark portion is also quiet useable.
1. It integrates very well with scala or python. 2. It's very easy to understand SQL interoperability. 3. Apache is way faster than the other competitive technologies. 4. The support from the Apache community is very huge for Spark. 5. Execution times are faster as compared to others. 6. There are a large number of forums available for Apache Spark. 7. The code availability for Apache Spark is simpler and easy to gain access to. 8. Many organizations use Apache Spark, so many solutions are available for existing applications.
Microsoft does its best to support Synapse. More and more articles are being added to the documentation, providing more useful information on best utilizing its features. The examples provided work well for basic knowledge, but more complex examples should be added to further assist in discovering the vast abilities that the system has.
Spark in comparison to similar technologies ends up being a one stop shop. You can achieve so much with this one framework instead of having to stitch and weave multiple technologies from the Hadoop stack, all while getting incredibility performance, minimal boilerplate, and getting the ability to write your application in the language of your choosing.
In comparing Azure Synapse to the Google BigQuery - the biggest highlight that I'd like to bring forward is Azure Synapse SQL leverages a scale-out architecture in order to distribute computational processing of data across multiple nodes whereas Google BigQuery only takes into account computation and storage.
Licensing fees is replaced with Azure subscription fee. No big saving there
More visibility into the Azure usage and cost
It can be used a hot storage and old data can be archived to data lake. Real time data integration is possible via external tables and Microsoft Power BI
