 Hi, this is your host Bharatiya and we have with us Thomas H founder of salt stack and senior stuff engineer and director of engineering at VMware Thomas is good to have you back on the show. That's great to be here. You folks recently announced Item project, which is created to kind of deliver infrastructure as data and kind of to simplify cloud management automation So that's what we're going to talk about today, but let's just start with some of the you know Origin story What is this project all about and why you folks created it? Item was a research project inside of inside of salt stack that we had just started to play around with right before the acquisition and Shortly after the acquisition A number of people inside of VMware started taking interest The idea was to or be is to look at the way that we manage infrastructure Especially cloud infrastructure just a little differently instead of looking at it purely through the perspective of Needing to write code that enforces the broader state of the cloud That everything should be more full circle And this is also something that we learned at salt stack when we were doing our our sec ops product that one of the big problems that you run into in security is that all these security products Scan, but they don't necessarily have the ability to fix or they can't fix in a big way And we wanted to apply that to more generic management as well The ability to look at exactly what's happening inside of a inside of an infrastructure inside of a cloud Pull all of that information back in and then immediately turn that into enforcement and action So that we take those insights But we're able to do what we need to do from an infrastructure management perspective Directly with those insights so that the overhead of authoring thousands of lines of code to maintain those infrastructures Hopefully can be significantly reduced now. Can you also talk about how different is the your approach? You know as Infrastructure as data versus what we talk a lot about these days, which is infrastructure as code So item itself is founded on the ability to deliver infrastructure as code when somebody goes to learn item They're going to be introduced to what it means to use item from an infrastructure as code perspective But the big difference is that the way the way that everything is compiled and executed under the hood Means that we can reduce everything down to completely structured raw data and it becomes much more malleable That is also one of the things that gives us kind of this full cycle what we want to be able to do is Look at a cloud infrastructure gather all of the information about what's going on convert that into into code that can be managed and viewed and really handled Through this data pipeline so that it becomes again just significantly easier to build everything So that the amount of the amount of actual code that needs to be interfaced with goes down What are the core component what technologies from salt that you brought in there and also it's very new But what are open source component are there that are part of it? Yeah, so inside of salt stack I If we go back to kind of the early days of salt stack salt itself was made to be highly modular and extremely Plugable and over the years I played around with a lot of those concepts of Plugability inside of software design and Produce something called pop or plug-in oriented programming And then I originally thought well, let's let's take some of those concepts inside of salts config management backplane and Rewrite them on top of plug-in oriented programming and that's where item that is a core system originally came from And we realized that all of a sudden we were able to do things with asynchronous programming that just wasn't available before We had been able to take a lot of constructs inside of salt in the way that Interdependencies are managed and the way that we're able to deal with really complex orchestration and make it more powerful more Flexible and even faster than what we had in some salt And so we realized that this would this would make sense as something that could be applied to cloud automation In early days of open servers so good all you had to worry about red hat canonical and Suza and you're done now There's so much open source, especially in the cloud and you feel so many new technologies came coming up and now You know item is there too. So how are also trying to like I mean, of course the younger doesn't need much You know you folks have all the muscles there just to to I mean what kind of challenges are you seeing there to Get people, you know made aware of the project Who is going to be a target audience and what is your plan for the project, you know? Because a crowded space. Yeah, I mean the the playing field as you've said has changed dramatically In the last ten years that used to be that the mere existence of open-source software was enough to get people excited about it but now as you say there is just a massive flood of Software that's out there a lot of what we want to do is make sure that we can illustrate to people that this is going to make a significant change in In how they're managing infrastructure We want to make sure that what we can deliver is really an order of magnitude better than what they're then what they're doing right now through this process of Discovery description of an infrastructure and turning that directly back into action Because again when we look out at all of the different pieces of software that are being used to automate the cloud and automate again a lot of things They are so very heavily dependent on The developers and particularly SREs to the extent that we've got something that we're starting to call the SRE crisis We just can't keep up with the number of SREs that we need to to run these infrastructures the scale Continues to get bigger and bigger and in doing so we just can't train up or have enough of these folks in house Now we've got to make we've got to lower the burden on the SREs so that they can get even more done with less And so that is the big differentiator that we're really going for since you brought up the point of SREs I also want to quickly talk about The rules are evolving, you know people can very easily define who it is, but then also What kind of things fall under SREs buckets some folks said hey sorry should know or DevOps should not even be your job title It's totally different thing. So first of all, how would you define if I ask SREs? What do you think their rules should be in today's world? So in large part The roles have shifted dramatically and especially if we come back and we look at kind of the evolution of the whole Space over the last I would say about 15 years since we have the original introduction of the DevOps concepts We we tend to recreate silos Over and over and over again the core idea behind DevOps was really about inter-team communication And the tooling that we built around DevOps when we go back a decade, you know, 10 years ago and I was Making salt we're really around those ideas of bringing together developers and operations professionals And in many respects it worked, right? We created we created development pipelines The DevOps the DevOps people who in many cases have turned into SREs They created these pipelines making it easy for developers to be able to commit code And then that code gets tested validated and deployed so that everything becomes a nice smooth experience but what it's evolved into as Happens virtually every time we try and do this Technologically is that we've begun to recreate those silos so that the SREs are becoming separated from the developers again And they're also separated from a lot of the other operations that need to happen around an infrastructure And in doing that separation means that we're ironically running into the same problems we used to have But we're also creating automation pipelines that are so much heavier than they used to be Used to be that you would deploy code many many times Day and you would be able to test it quickly, but many of these pipelines now are Loaded really really slow And we and we need to rethink how we're approaching that problem Thomas Thank you so much for taking time out today and of course talk about this project I also share some insights about the whole evolution of sREs and you know the challenges that are there As you said the project is new it will mature so I would love to see you again as you know We see more you know releases of the project or more updates there, but I really appreciate your time today. Thank you Yep, thanks for having me