Apache Cassandra Apache Cassandra World Party | April 2021, 28 Apr 2021

Previous Meeting

⏯

youtube image

►

From YouTube: Understanding Cassandra | Apache Cassandra World Party 2021

Description

Presentation from the first Cassandra World Party on April 28, 2021. Learn more about Apache Cassandra at https://cassandra.apache.org/.

Pradeep Gopal of Target presents “Understanding Cassandra” touching on the successful Cassandra business cases and the learnings - especially on what to look for when you are setting up Cassandra.

About the Presenter
Pradeep is a principal engineer at Target who loves to work on Data problems. Prior to Target, he's worked on Cassandra and related technology to handle massive amounts of data in IoT, education, and the social welfare space. At Target, he works on the usage of Cassandra in inventory management and visibility.

Apache Cassandra is an open source NoSQL distributed database trusted by thousands of companies for scalability and high availability without compromising performance. Linear scalability and proven fault-tolerance on commodity hardware or cloud infrastructure make it the perfect platform for mission-critical data. Get started: https://cassandra.apache.org/quickstart/

The Apache Cassandra World Party 4.0 was a virtual, one-day celebration on Wednesday, April 28, 2021 in preparation for the launch of 4.0, the most stable database release ever shipped.

Join the Cassandra community: https://cassandra.apache.org/community/

#ApacheCassandra #database #NoSQL

A

I'm excited to introduce our next speaker, who is pradeep gopal he's a principal engineer at target. uh Work loves to work on data problems, he's worked with cassandra and related technology to handle massive amounts of data with iot education, social work in the social welfare space, with target he's worked in particular about with cassandra on inventory management and visibility.

A

um He's I I have a special.

B

Place in my heart for.

A

Target having worked with them in the past and sat in more work, more rooms with those guys and so uh without further ado. Let's, let's get to see what what uh pradeep has in store for us pretty are you on.

B

Hey yeah, hey good morning, good evening, hope you're all fine and safe yeah. Can we can you get the slides up?

B

Hey uh I'll, just briefly touch up on couple of things about why we choose cassandra not necessarily a target even before a joint target uh and uh what we learned and what are the set of use cases that we built and what are the sort of key takeaways.

B

Sure so why we choose standard answer to that question is because there are tons of data that you would like to slice, dice and store, and that is true for our case and seven years back when we start looking at data store, we needed something that was fast to read and write and no sql db is where the best to uh get going and cassandra was the new kid on the block and does he want to give it a try?

B

And that's why I started it was easy to learn, but I'll tell you more about what we learned and how it happened. It appeared to easy to learn: that's that's what it uh essentially meant. So it was like. You know, create a set of partitions. You can pick your partition keys store as much as data you want and that's how we all uh started with that next slide.

B

So how did we learn kassandra, of course a lot of resources available, but it is not simple or simple to use or implement just like any framework into you need to have the experience you need to spend ads on. It's not like. You read a book on cassandra and be able to try to implement something and.

A

B

At the right side, the the curve that you see is the ramp up time on cassandra seems to be less. It is steep, but it is less, but once you reach the point where you think you know everything that you know about cassandra, then it takes a good amount of time to master and the ramp down. Time is little longer than also what the ramp of time. Eventually is. That's what we learned so yeah. So what does cassandra allow us to do? Basically, it allowed us to scale data at a real, real big scale.

B

Right taught us how to store and uh large amounts of data very efficiently, and we are able to build the data system around the architecture, not the other way around it. Typically, what you do is you have a data base rdbms typically, and you try to build architecture, there's cases where I've seen where you cannot do something in the architecture, just because your database cannot support that right and your system is prepared to basically uh connect to various new business cases and use cases at actually gollum a set of use cases that we worked on.

B

What I, where I used for the scale, was, was iot platform. It was typically a smart water management system used to get data from census, 10 million to 100 billion per day, and there's not only that. We also got archived data, which used to be like uh we used to get around a billion events in a span of day or two, and this data has to be sliced and diced across the time series, and we should also be able to do the same thing of course hierarchy.

B

For example, when I say time series I would like to get a aggregation of lecture average rainfall uh in an hour across a given location and set off the average across the overall geographical location, with all the sub things, and one of the key takeaways that we have from this was relevancy of data is completely perfectly acceptable, don't be afraid to redesign your desktop because many times we actually have to throw away the entire thing that we have built and rebuild it. And what also? What else else is cassandra right?

B

It helps you design and write your pipelines so that you can handle the data in a better way right and design your design and tune your clusters for rights, reads and shelf life of the data, so something that data for one year old data. You may not really need to have it in your current cluster. You probably can do it in another cluster or another key space, another table and restrict your partition size and choose your primary key wisely, and this is something very critical, because this is what we found.

B

Having restricting your partition size to 100 to 200 mb right, I guess perfect, fantastic performance. Again, that's my opinion. This is my experience right. So what are nice little things about uh cassandra one? Is? It can be a fantastic beast that can right once you can. You know how to tame it. It is restrictive, yes, but it also opens up various options to represent the same data for your uh business needs right.

B

The storage design is driven by the architecture, like, I said, not the other way around and each time that I have used cassandra for a new use case or something I've always discovered something. I've also learned something new about cassandra.

B

Finally, I would like to thank uh cassandra, patrick and the community, and uh every time we had a problem, we always used to go to their site. Look at the go through the sessions go to the slides that are there there right to learn about cassandra yeah. That's all. I have I'd be very happy to have any questions.

A

I'm pretty curious about- and I told him before this- I'm curious about how they're doing inventory systems with cassandra, because man there's so many inventory systems out there and sometimes it's hard to to group all of those things in one place so that you can it. I was talking to some other companies that they were like if I click and come into the store, I really need to make sure that that thing is in the store and and cassandra is a great place to centralize that data.

A

So thank you so much pretty for preparing and presenting today.