Hi, I'm Anna Hoffman and welcome back to Data Exposed. Today I'm joined by Robby from the SQL product team. Robby, thanks so much for coming on the show. >> Hey Anna, thanks for this opportunity. It is always good to talk about SQL Hyperscale. >> Awesome. We are super excited to talk about Azure SQL Hyperscale. I always am. Uh and today I really wanted to bring you on cuz you were a big part of this new Hyperscale benchmark that we did. So, can you tell us a little bit about like what do we do a benchmark and who actually does the benchmark and what does it even mean? And then we get into the results? >> Yes, definitely. So, this was a benchmark commissioned by Microsoft and an independent vendor did this. So, it is an independent study. And what they did here is they were simulating HammerDB. That's a testing benchmark database. And they tried to simulate order processing system like an e-commerce and trying to see that how many orders at a high concurrency actually this database can perform. So, they compared Hyperscale against Amazon Aurora. So, that's what this benchmark is about. >> Okay, cool. And then this benchmark got printed and released by this third party like you were saying, Principal Technology. We'll put a link in the description for folks if they want to read it themselves or actually the link is right here. But okay, so we did this benchmark against us and Aurora and what were the results? >> Results were really uh amazing as usual. We know Hyperscale how great it is. And so, what we saw there is Hyperscale actually did nearly 52% better comparing to Aurora when we were talking about raw performance, which is great. So, we did it at different concurrency. There we see 64 users, 96, 128. So, we are seeing that as we scale the concurrency numbers also Hyperscale is still performing way better than AWS Aurora there. >> Okay, cool. That's great. And then I assume we're trying much as we can do like a apples to apples. So, when this says 16 Vcore, I'm not really familiar with Aurora sizes, but like we decide or or the third party decides that that's the similar size as a 16 Vcore. >> Yes, exactly. So, to be really fair, when we are picking 16 Vcores equally be Hyperscale there, we pick a extremely similar. I mean, huh Principal Technologies picked a similar instance there, which is DBR7X. I mean, that's their internal information on how do they pick these instances, but we want to be really fair because that makes it valuable for customers. >> Okay, amazing. Cool. Okay, so performance, great, check, Hyperscale's awesome. Love to see that. Okay, but it's not just about performance, right? Because as, you know, as a customer, I can pay for more compute, and then I can get more performance in theory, I guess. But, it's also about price, right? So, how do we compare there? >> Exactly. So, that's very, very interesting line which you said there. So, you can always throw more compute and get more performance. But, as a customer, I definitely want to understand that can I drive the best perf at the right performance um price per dollar as well. So, what they have done is they have uh divided number of orders by the total computer hours in a month to understand how we are actually faring in terms of price per uh dollar or perf per dollar. So, we are there doing again extremely well, 51% better compared to AWS Aurora here. So, 51% better when we are doing perf per dollar to process number of orders as well, which is great news for Hyperscale. >> Okay, awesome. That's great. I obviously am a big fan of Azure Single Database Hyperscale. Hopefully, more folks are also a fan of Hyperscale. Now, one thing I wanted to touch on as we kind of think about wrapping this episode, we said better performance, better price performance than Aurora. Um when it comes to Azure SQL database Hyperscale, are there other things that people should consider uh before they use it? You know, like is it only for big databases or is there something something else they need to know? Or is it like nope, it's better better performance, better price, just let it rip. What do you think? >> [laughter] >> So yeah, that's that's great question. So customers we want to meet customers based on whatever their needs are. We want to give them that flexibility. So at Hyperscale it is not just for large databases. You can start small, subcore, we have serverless. You want to pack databases, we have elastic DB pools as well. And you have that ability to scale from a small few MB database to at least 128 terabyte. So you have this huge flexibility. You can grow your app, you can grow on Hyperscale and we are there to all along this journey. So definitely you want the customers to start, pick the right flavor, and uh scale with us. >> Okay, awesome. Well, Ravi, uh thanks so much for coming on the show. I learned a lot. Uh to our viewers, go check out the benchmark for yourself, read it yourself, it's out there, it's done by a third party, not us. Uh so you don't just have to take our word for it. Uh but hopefully you will take our word for it also because we think Azure SQL database Hyperscale is great uh for your next application or for your upcoming modernization and innovation. Uh so with that, Ravi, thanks so much for coming on the show. To our viewers, uh thanks for tuning in. Leave us a comment, let us know what you think. Uh we'll put this link in the description so you can learn more and we hope to see you next time on Data Exposed.