.

News

22. 11. 2022

Accelerating Data Solution Migration to the Cloud

Whether your legacy solution uses enterprise ETL tool or hard-coded PL/SQL procedures, Adele provides comprehensive list of the connectors to both options. Rely fully on the Adele's capabilities to extract knowledge and understand the logic of your data solution.

Listen to the podcast (EN)

Speaker: Luboslav Gabal

Subscribe to Adastra Podcast

None

A podcast that inspires, advises and that you simply have to listen to if you are serious about digitization.
Don't just buy an IT solution - buy the result, generate higher sales, reduce risks and have less worries and more time.

Podcast is recorded in Czech language.

Read the podcast as an interview

Typical cloud migration challenges in data management domain

The journey itself might be still in front of you, or you are in the middle of migration. But it's good to know that migration itself is not a one off activity, it's a process. So there is a lot of effort related to pre-migration activities, like understanding the complexity of the system that you are about to migrate. Profiling of the data, profiling and understanding of the data pipelines complexity. Of course, most of the effort is then related to the migration itself. So reimplementing the actual data pipelines or ETL processes, orchestration migration. In your current solution, you most likely have some workflow management which is managing the order of your data pipelines. And after the actual migration, of course, there is a lot of effort related to making sure that the solution in the target technology is the right one, is the correct one. So lots of effort related to actual reconciliation of the target solution compared to the old one, a lot of end users testing and performance testing and optimization, obviously.

Typical obstacles related to running your BI ecosystem in the cloud and perceived downsides of cloud-based data solutions

Of course, these are unexpected costs and very often unexpected costs of the migration itself. When calculating business case, one very often thinks about the cost of the operating of the solution in the cloud. But the efforts related to migration itself are very often underestimated. The complexity of the system dependencies between the data consumers and data providers are very often underestimated. When doing the migration, the natural decision is to utilize native technologies provided by the cloud provider. This then naturally leads to vendor lock in, when these services are not always inter-exchangeable between the cloud providers. And then the usual suspects like GDPR compliance, data security and data governance that is not strictly related only to cloud based solutions, but generally to any data solution, either in legacy or cloud world.

Adressing the challenges of all three phases

Fortunately, we do have a solution which is addressing all three phases of the migration, pre-migration activities, migration itself and post migration. The solution is called ADELE, and we developed it especially for the purpose of accelerating and speeding up the actual migration to the cloud.

Understanding the legacy  and speeding up the migration

So let's say that you are running your legacy data warehouse in Oracle, ETLs in hardcoded PLSQL procedures. The solution was built 20 years ago and as a typical legacy solution, it bears plenty of workarounds, almost no documentation. The people that were working on it are either working in different department or they left already. And now you‘ve decided that it's the right time to migrate, to move the solution to Microsoft Azure to utilize data Azure data factory as the ETL platform. So what ADELE does basically is extraction and harvesting of the existing legacy metadata, either from metadata driven ETL solutions or from hardcoded solutions like the one that I just mentioned, hardcoded PLSQL procedures. It collects the metadata, puts them into one common metadata repository. Of course, keeping the dependencies between ETL jobs and really understanding also the data lineage itself and then automatically generates the data pipelines in the target solution. In this case, it would be Microsoft Azure data factory. What this allows you to do is of course understanding the legacy solution, understanding the complexity of the solution. It allows you and suggests you to simplify and declutter the solution. What that really means. Often when you are analyzing existing legacy ETL jobs and data pipelines, you discover plenty of unnecessary legacy stuff like ETL jobs that are just running there, it doesn't have data consumers, but probably no one had a courage to decommission it, due to the fact that no one really knows what it does and who is consuming the data. When you are doing the migration, the very often decision is to do also business redesign of the solution along the way. Should it be data warehouse or data mart or big data lake, you probably want to extend it with additional entities. You want to decommission certain parts of the solution, decommission certain data mart. And this is also something that Adele fully supports, redesign of the solution. And last but not least, as I already mentioned, replatforming to target technology. Should it be Amazon Web Services, Azure or Snowflake Google, the typical suspects when it comes to target cloud based technologies. Adele is really technology independent, technology agnostic. So we are supporting plethora of the legacy technologies like you can see on the left side. When it comes to the target technologies, we support the major ones, Amazon, Azure, Snowflake, Google naturally, with different deployment strategies and different tool stacks in these target technologies.

Extensive data lineage and impact analysis capabilities

And your data journey doesn't usually end with migration itself. What I often say is that data warehouses and reports and data marts are the solutions that are never finished. There is always a plethora of new business requirements coming onto the plate necessary to be implemented. And for that purpose, you usually want to have a very good overview of your solution, whether it is in legacy technology or a cloud based solution. Well, one of the side effects of using Adele for the migration itself is that you have metadata all in one place, in one central repository, which allows not only extensive data lineage and impact analysis, but also very quick further replatforming, which is exactly addressing one of the fears, and this is the vendor lock in. So shall you decide in three years, five years, to replatform your data solution once again, meta data are there if they are kept up to date, which is really easy to do with Adele, you are ready to re-migrate once again with much less effort compared to the first migration. So thank you very much for your attention, and shall you find yourself in shoes of someone responsible for project of migration of data solutions to the cloud, you are free to reach out to us, we are happy to assist you. Thank you and have a nice day.

Discover more about Adele: www.adele.pro

Interested in working for Adastra?


Take a look at our current job vaccancies.

Do you like Adastra but feel that none of our career offers fits you? Feel free to contact Dasa Noskovicova via LinkedIn.