"Welcome back to that of his podcast sunless here with you. Great to have you back and it's another reinvent update show show today give you a bit of a break to catch up before we have venevision keynote Tamar some really cool updates In broad brought areas particularly in compute and civilised but also some Datta ones that are very cool and a few Buddhist wants. So let's get started. Firstly oedipus Lambda announces the provision can currency. So this is a feature that gives you Greg Control of performance of your civilised application at any scale functions using provision concurrently execute with a consistent started latency so this is really useful if you go as interactive mobile aware beckons APR Beckons Latency Sensitive Marker Services etc.. Essentially what you you can do now is simply set. The expected currency on any version or alias. A function from then on and he was in chose. The function begins executing your code within double-digit mealy seconds of being invoked so this really helps it use cases we've got an APR that might be hit periodically. you want to reduce that startup. Stop talking for that. Initially invocation people have little. Toss that run to keep. Keep a lambda around by pink from time to time. Now you can use these provisions currency capability to Kinda for that particular situation now. The good news is you can also use application orders. Scaling can automatically configure the requiring currency for these functions. You can do this using target tracking and it should yield. sculling policies will says Sam and the SAM's Eli can support provision currency and Salsa integrated into Kurd deplores. We'll take it really do it. Entertain speaking of capacity Amazon. ACS capacity providers available. This is a very cool new capability that will also include a couple of other features will speak about in a moment. Essentially these are a new way to manage compute capacity for containers and it allows the applications. That font it's it's requirements for how eases a capacity so were these capacity providers. You can define very flexible rules about how these containerize workloads run on different types of compute capacity and and you can manage the scheduling of that capacity really focused on the availability scall- ability and the cost of running tasks and services on ATS. So in the past you could choose the talk which was either. EC Two or eight. You could also use placement constraints and placement strategies to control where the toss would execute however you could only use the capacity that was running. You couldn't sort of anticipate the capacity now. These capacity providers work with both easy to end. Fog It so with E C Two there are based on easy to order scaling in groups and the capacity provided manages a scanning of the order scaling group through the ECS CLAWSF DOT org scowling which is located. I'll talk about that in a moment. Also also if you're deploying on filegate you can also use fog faggot spot capacity providers and we'll talk about five spot in the moment as well and you can split between multiple capacity provides. So if you think about the way you balanced your work leads the way you manage your underlying infrastructure gives you more choice and Alomar automation in terms of polio class teasing. So I kind of teased the Knicks to Update so Amazon EC Cluster Order Scaling is now available so this means as your easiest classes on Itchy automatically skull based upon your scaling rules and of course he sees the capacity provided construct that we spoke about so so in the past. It couldn't manage the group you had to do it manually no longer nine more. You can integrate the whole thing together. Also as I mentioned fog at spot is now available. She can save up to seventy seventy percent for fault tolerant applications so essentially if you have an application way the Container goes away and comes back very quickly. It doesn't really met US things like big data Siasi today's batch processing etc it is not integrated into guy."