Please share with the community what you think needs improvement with Microsoft Azure Synapse Analytics.
What are its weaknesses? What would you like to see changed in a future version?
When I was trying to link services to an SFTP site it was not able to do all the possible encryption that I needed. They can improve by adding more encryption options.
Indicating what areas need improvement in this solution is a difficult question because the organizations that I am working for are really new in this area. However, an even better more simple interface, or perhaps an extension of a connector app store solution, would be helpful. For example, if your organization, has an electronic client dossier and you want to connect that information system in real-time towards Power BI, you just go to the app store and see if there is an ONS connector, introduce it, set it up, and go from there. In an upcoming release, it would be beneficial to have the ability to have dashboards on mobile devices to receive a better experience for the user.
For improvement, I would like to see more connectors and better integration for Hadoop. There are currently limitations for Hadoop on-prem. I'd like to see better integration and more flexibility.
They should provide a less expensive version with a smaller setup for small businesses. Currently, its price is quite high for entry-level or small businesses. In terms of integration, new connectors are always welcomed.
The solution needs to improve its integration capabilities. It needs better integration with SQL and other databases. The initial setup is complex. The product needs to implement more machine learning and artificial intelligence. If they had more IoT and streaming capabilities, it would be helpful. The product could use more analytics tools. It would be helpful if we could use the product with less coding.
What I would like to see is a possibility for time travel on the data. For instance, a financial department would like to do an analysis of the data set as it was for the first of the month while another department wants to run the analysis about how things went yesterday or last week. They have different starting points for the queries. I've seen Snowflake handle tasks such as these. It's got a feature called time shuffling, which can show me what's there on a certain variety of dates. It's very convenient. It would be great if we could query the data of the day as well as in the data warehouse. They should have scalable memory processor options for the data warehouse. While the solution is flexible, sometimes this works against the user.
We've only recently starting testing this solution and it would be helpful if the documentation was better or if the company could provide some kind of presentation to help us understand how the tool works and what's required to optimize use of the solution.
The macro functions, though useful, are not totally user-friendly. Some people have difficulties in learning them. If the project keeps growing we will need to link some of the documents to the data warehouse. However, at that time, I'm not sure about the possibility of keeping the same systems that we are using right now. We may need some kind of migration. I cannot say for sure, as I am not an expert. We'll need advice on that. The initial setup is not straightforward. The solutions not user-friendly and does require training of some sort. The solution requires a lot of the memory of the different devices.
More integration is needed to improve the product for the future. There needs to be more flexibility when developing applications with this product.
There aren't any features that are really lacking in the solution. We don't really have any issues with it in its current form. When we used the on-premises deployment model, we had data latency issues and suffered from page performance problems. However, since we've moved to the cloud, we haven't had these problems. The initial setup has a bit of a learning curve.
Unfortunately, we have had some issues with the dashboard reporting. Sometimes, the data for specific periods would just appear blank on the dashboard. To investigate this, we worked with a Microsoft incident agent and it turned out to be a result of bugs in the platform when dealing with specific types of queries in Azure Data Factory. We have also encountered some bugs regarding incompatibility with certain versions of Power BI. In the past, Power BI used to integrate nicely with Azure Analysis Services which we used as a bridge between Power BI and Synapse Analytics, but they have since discontinued the features that worked so well for us. Apart from fixing bugs, I would also suggest: * The learning features and documentation could be improved and expanded on. * The calculation and forecasting tools could be made more clear and easy to use. * The price could be lower.
I think that there should be better governance. Actual pipeline level security, if you create a pipeline within the tool, because it's a one-stop-shop for everything, you control who has access to what, and that goes for anything, whether it's an internal or an externally managed table. I think there needs to be an improvement with support for external tables supporting Spark, but further support of Lakehouse to make it a true data warehousing product where you have this true separation of computes and you have a separation of storage. I think it kind of dips its toes in there, but in order to achieve, I think with a vision that Microsoft is doing, trying to get at to replace Bolt, but compete with Snowflake and Databricks at the same time, it needs to have those features on. In the next release, I would like to see Lakehouse and better governance.
The performance and data consistency need to be improved. Support for PSQL functionality should be included, as it is something that is missing right now.
I am pretty sure that there are areas that need improvement but I just can think of them off the top of my head.
I am very sure that there are areas in need of improvement, but I can't recall what they are off the top of my head.
After only using the solution for three months, we haven't yet really gotten into every corner of its capabilities. It's hard to say what might be missing in terms of features. It would be ideal if the solution could be better used intuitively by the staff without having a great deal of training. An organization should be able to adopt and operationalize it throughout their company with minimal training.
I am a researcher. For people to be able to research a solution, there should be at least a free trial. Just advertising a product or saying that this product is better doesn't work. I would strongly recommend providing a lot of free trials and trainings. This will also help Microsoft in having more users or customers. Oracle provides some free trials. You can just go for a free trial and use your database online, which is very good.
Technical support needs to improve. Microsoft improves all of the time, but there are some things where different vendors offer data warehousing and data governance that you can build and workaround. It would be beneficial to take the top vendors and identify some kind of straightforward action to work with them. Instead of having to employ a separate vendor tool to be able to move this, it would be nice to be able to go through Microsoft.
What I would like to see more and more of in Azure is its support for IoT and streaming media information. It needs strong support for social media, internet data, and native support for NoSQL. At this time, it works very well with the structured data Stability could be improved. Technical support is very good, but could also be improved.
Microsoft should develop an interface to make it easier to shift from on-premise to the cloud. There needs to be a lift and shift tool to move on-premise, IMS high-volume data to synopsis sequence data on the cloud.
I would like to see them provide the ingestion of images.
There are features coming int he next few quarters that will be helpful. Soon, Power BI will be directly integrated into Azure. We need to have some Spark tools also available so we can directly select customers and don't need to install everything. There will be features added that relate to application development. There's hopefully going to be more flexibility with the XML. Right now, for example, Data Warehouse is not able to give XML files and your file put is not correct. The feature will hopefully allow us to read XML. The performance needs to improve in future releases. We're hoping that Microsoft will add integration with the Amazon AWS platform.
This solution needs to have query caching so that if the same query is run and the results are available, it will return the data from the cache without having to re-run the query. As it is now, you still have to use compute for each query. We would like to have integration with the data lake so that we can avoid retrieving the data from it, just to store it in Synapse before we can use it. From an architectural perspective, avoiding this replication would be good.
I would like my team to be able to build pipelines that integrate with the Azure Data Factory. Some of the best-in-class ETL products that are available in the market, such as Informatica ETL, are far more mature than this solution.
The major concern is that I have a hard time with having to version control the data warehouse all the time. As it is now, I have to open all of the ports and push everything onto the server. I would like to see version control implemented into the data warehouse. This would make the tool perfect.
The configuration for things like high-availability could be more user-friendly for non-technical users.
So far, with the current SQL that I'm using compared to what we've previously used, I'm happy with it. I wouldn't necessarily point out anything that needs improvement. From my point of view from what we have implemented recently, I'm happy with the solution, I'm happy with the performance. I'm not entirely happy with the billing model. I'm not entirely happy with how the enterprise services are pretty expensive, but that's about it.
With respect to what needs to be improved, concurrent connectivity has some limitations.
There are several specific features and functions available on the on-premises version that do not work with the cloud version. There are many issues regarding migration. They are building it more and more, but if you have the same version on local you cannot just go to the cloud because some of the functionalities don't work yet. You would have to re-code everything, and maybe even do it differently. In terms of documentation, they can make it easier or start with some advising or instruction. As it is now, it's like making a puzzle. Technical support needs some improvement. They will need to include some features for the Cloud version. They can make subsets of the database so you can schedule the marking department to work only between nine to five, but ETL can run at night. With a subset database, you can cut costs, similar to Snowflake whose price is a lot lower.
As far as improvements to the product, I think the Data Factory integration should be enhanced to be properly implemented because right now it is still lagging behind other vendors like Informatica. We ended up having to use Data Factory because we don't have any expertise in Informatica. But I want to see Data Factory get better overall. An enhancement I would like to see is in data loading for Data Factory. Right now we are seeing a little bit of a performance lapse in the cloud server. Specifically, this happens between Azure and our clients maybe because we are connecting to SQL Server Management Studio. There is a slight but noticeable delay. I do not know if it has to do with Azure and whether it is their problem or our problem. I can't complain much about Data Factory, but because I used to use SSIS (SQL Server Integration Services) which was the prior on-prem version, I know what capabilities it is missing in comparison. They changed it over to Data Factory, but by comparison, Data Factory has a long way to go to reach the level of SSIS as a product.
It would be of interest to improve things like the web service integration and availability in terms of being easy to create internal web services in the database.
The initial setup process needs improvement. When you're moving to the cloud it takes a bit of time. It would be great if they could implement something that would make it faster. It takes a bit of time but once you're done, you're done. The simplicity is good, but it should be that anyone can use it. You shouldn't have to be technical because, with the way it is now, only a technical person who has had training can set it up. If you have a dashboard that anyone can use, it would be easier.
There are several things that I would like to see improved with this solution, including: * The price of this solution could be improved. * I would like to see better integration with Active Directory, because we have had problems, and we still do. * I expected that Microsoft would have had something out of the box for Office 365, but they do not. * It is difficult to configure, especially if you have multiple domains. * I would like to see a simpler interface for the Azure portal because it is kind of complicated. * It would be nice to have better monitoring of the usage of the data inside Azure SQL. What we don't use could be achieved in cloud storage. This would allow us to cut down the price. * Better customization would improve this solution.