18 Burst results for "Kuban Eddie"

Managing Research Needs Using Kubernetes with Bob Killen

This Week in Machine Learning & AI

11:34 min | 1 year ago

Managing Research Needs Using Kubernetes with Bob Killen

"I am here at q Kahn in San Diego. And I'm with Bob Dylan. Bob Is a research cloud administrator at the university. -versity of Michigan Bob. Welcome to the PODCAST. Thank you awesome so We met yesterday you were on the AM L. Media and analysts panel talking a little bit about your experiences using communities to support the researchers there At the University of Michigan Gin. And you're also one of the other rose plsy are co chair of the CNC FS research User Group So really interested in hearing a little bit about your experiences with. Sounds like you work with Kuban. Eddie's kind of to support a broad portfolio of applications not just MLA. I we run a wide variety of applications of both animal focused like keep blow but we have a whole slew of other supporting services that are researchers Consume Joey on in our larger. HPC Cluster. Tell us a little bit about your background. How did you get started working with communities? I moved over to artistes Microbe advanced research competing technology services about three years ago but before that I worked for the University of Michigan Hospital and in Twenty fifteen we we started going down the container out for a lot of our clinical workflows. This is sort of before docker was a thing So we are predominantly like Alexi containers. And we're first looking at communities to orchestrate that but raise wasn't mature enough of the time and we went For us from there we Dan Shift back and forth and also start supporting our research workload serve alongside our clinical workflows on there. They sort of saw you know some of the benefits of containerization and this WanNa take advantage of the broader service that we were offering From there I was hired by artists to sort of build out the same something but be completely. Research focused as well as like managing of virtualization system and communities has really exploded when it comes to the the broader management adoption of containers and We have things like that. Keep sputtering in super hub that just integrates directly with it in from. There's just inside the cube. spiner cube sponsor sponsor in reporter wanting to hub at. It can spend a notebook as a container itself itself. Can you maybe talk a little bit about the various use cases that you're supporting what researchers doing. Oh man it's a a pretty broad stroke spectrum on top of reneges. We support a bunch of social science stuff. Various databases that are consumed By in other applications and people running workloads in our large H. B. C. Cluster And we do support We have a significantly large a jupiter environment a lot for classes and other researchers that are spinning stuff up bioinformatics physics Think those are most most of the big ones okay. And so is the primary user experience among your researchers the Jupiter notebook and the ability to spun off containers from the notebook amendment. Yup over time. We've seen a gradual shift of people you know sort of moving away from the classic. HPC to style system. Where you would log into a sage along node? And you'd run your system you something up and running off now. They're sort of an expectation for art to have the Jupiter Notebook or have some other sort of science gateway for the user to consume. It's it's a lot friendlier than than having to dig in and right like a bash script it to automate some of the suffer you one of the things that's different between the classical. HPCC scale up systems and the more distributed environments that We see now and that are common With Kuban. Eddie's is that in order to take advantage of I've these distributed environments the users in your case researchers have to know about them and kind of know how to to to use them in a lot of cases right there you know code Oh to take advantage of distributed computing. Are you researchers doing that or are you. Are there things that you've put in place that provide abstractions so that they don't have to Think about that We do have some abstractions and play that Make it much easier argo workflows and things of that. Nature have simplify things greatly for people then of course empower users. That you know will will just want to dive in and do everything themselves. But in general we have a pretty good suite of tools and libraries to make it easy for them. The social scientists isn't kind of the target user for a cube flow yet they're using communities Q.. Flow in your case. It sounds like the ARGO. WORKFLOWS you know our lot. You didn't say that necessarily the social scientists Mantis were using the ARGO workflows in Q.. Flow but I'm I'm curious to hear a little bit about how cube flow in particular is used in this environment. Cute flow it. So we don't have any social science people using cube flow The cute flow stuff is mostly by a small subset of people. And Right. Now it's a more experimentation stage But the the aspects of it released sort of make a special like the model life cycle of things Significantly easier for people and we we sort of see much. More people shifting. Adopting it going forward maybe talk a little bit about that user experience. What do they have to do to get their apps up running into keep environment? They don't have to do too much they can do. Most of it from the Jupiter notebooks that Cupola spins up in manages for them and then there is I forget the the Argo are the flow Cli tool but that allows them to create some workflows pretty easy to. There's there's some other things being built that allows allows for better like designing and building better pipelines in More consumable workflows. Maybe talking generally. What from the your participation the patient in this panel? I got the impression that you're pretty excited about Kuban as as a platform generally for these types of research workloads. Maybe talk a little. What about you know why you're excited about it? Sure communities provides a lot of like proper abstractions for just managing a variety Heidi of workloads and those things that it cannot handle it offers the extension points to easily extend and augment so like grenades it self does not have any capability of running a like standard. MPI job or a gang scheduled job but because the extension points are there. It's easy enough to who sort of your own schedule or to handle that. And it's just like a little. You know one variable change than us that scheduler to spin up and manage that workload and and by using carbonates itself and the extension mechanism. You gain the accessibility of using everything else. That's being developed on top of whereas in the past you know again going back to the classic. HP system there wasn't any of these other this other tooling API's really do any of this stuff. So now we can get. You can do a lot more such as like eventing and triggering different workflows off. Different things happening within the system We're definitely seeing them. A much more broader adoption of things like data streaming an link in this stuff. You can really do well on a classic. HP system but it it works in fits quite well on top of communities and you can integrate it with a whole slew of other things that are being built on top of it. Are you supporting putting users. That are doing things like building out their own schedulers or are these things at the broader community is doing and they can just kind of take off the shelf and take advantage of right now. The it's mostly from the broader community and our users can then take that off the shelf cute flow itself has a NPR operator built into it. So that makes it easier for their the people consuming Q.. Flow to spin up in. MPI Job And then there's other things like a volcano which are trying to offer much more of the classic a batch computing things that we'd find. HPC for communities directly. Joe Down a little bit deeper into that. What is volcanoes? Specific volcano is a Open source project that is trying to be sort of the classic PC scheduler and workload managers thinking like slim but built for Kuban Eddie's so it is offering things like a fair share cues backfill as well as offering things like you know being able to support gang the jobs and things of that dream So the idea there Being that as a A research cloud administrator. You've got some pool of resources. You've got some pool of people that want to take advantage of these resources. How do you Fairly are consistently give them access to the resources and so oh my fair share of being one idea there gang scheduling is more. Like I've got these five things that need to go at the same time. Exactly how do I ensure that they're happening in concert with one another and then like if a worker happens to die you don't want to necessarily kill the entire job whereas you know if so the controller dies you want to then you know kill and Req- a Q. The entire job for us in for Sort of the larger sites that are still primarily a on prem focused on like bare metal you know we have a finite finite amount of resources. We aren't necessarily bursting up to the cloud So being able to backfill and set priorities on things is very useful for us as well as you know cues For us is like we want to backfill with jobs from students. They might be able to run those for for free but if we have a researcher that's doing something when we want to have those take priority sort of bump off those lower party jobs we think about in the context of machine learning Some of these. He's jobs Like network training deep learning Training they can take many days or weeks But that's not a new thing in the context of HVAC are the workloads that you're tending to support also kind of these long running Jobs but but not necessarily deep learning training we have we support again being a That handles researcher needs for the entire university. We have a little bit of everything and most of our jobs would probably fit into that category Where they might take a week to complete Or something of that nature I think are Maxwell time is twenty one days and what can you say what that is. I don't remember off the top of my head But we've had several that are dislike cranking through you know terabytes and terabytes of data And if they have something that just hits the wall time you know. It's it's unfortunate for the Kill it off. Also the this maxwell time isn't the biggest job that you've seen the limit that you yes yes all right the thing that runs logging in that we're going to give him the Nice Nice warm. Not Nice

Researcher Kuban Eddie Kuban Argo Michigan Bob Dylan Administrator Q Kahn BOB HP San Diego Am L. Media University Of Michigan Gin Joey Alexi Reporter Maxwell
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

02:02 min | 1 year ago

"kuban eddie" Discussed on Software Engineering Daily

"He hosts the Kuban Eddie's podcast from Google an excellent show about recent developments, and evergreen concepts within the world of coober netties, I use the Kuban Eddie's podcast to catch up with some concepts in the cougar Netease. Ecosystem prior to the most recent cube con. So if you are going to attend a coupon, I highly recommend binge listening to the Kuban Eddie's podcast prior to Google. Adam worked at AWS for three years and Microsoft for twelve years. He has seen each of the major cloud providers up close, and he has a deep awareness for how each company thinks Adam was fantastic. Guests, we had a great conversation about the cloud native landscape, the world of podcasting and developer marketing, which may not sound interesting, but you are being marketed to constantly and it's growing and. Competitive and fascinating landscape. And you as a developer, who is going to be buying tools, as much as building them should be aware of how marketing in the software landscape works. Speaking of marketing, a few announcements about upcoming things in the software engineering, daily landscape find collapses, the company. I'm building if you have a cool project that you're working on, I would love to see it posted on find collapse, and you can find some collaborators to work with on your project. I check out every project that gets posted to find collapse. And I've been interviewing people from some of these projects on the find collapse podcast. I'm attending some conferences in the near future. Data dog dash July sixteenth seventeenth in New York, and the open core summit which I'm actually going to be emceeing that'll be September nineteenth and twentieth in San Francisco. We are hiring two interns for software engineering and business. Development, if you're interested in either of these positions..

Kuban Eddie Adam Google developer AWS Microsoft San Francisco New York twelve years three years
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

04:14 min | 2 years ago

"kuban eddie" Discussed on Software Engineering Daily

"And I think also tensor flow and airflow and red is, and so in it's based on sort of what's a good fit for coober Netease plus, what is their sort of immediate demand for in had dupe isn't yet, you know, it isn't a. Fit for Cooper, nutty. So I think that hasn't that is sort of at the level where we're integrating with it, but not really running it spark is and has been a very good fit. And so we have native support for spark and Kuban Eddie's is a native spit scheduler for spark. And then clustered databases obviously are natural fit or customer data stores are natural fit for Kuban Eddie's, but then also things like my sequel cockroach db and things that are Maria DB kind of more traditional kind of databases as well. And then of course, red is elastic search. These are things that people tend to use very commonly with what they're building on. Kuban is I think over time we may find that Kuban Eddie's expands to be a more universal platform that certainly come a long way already. It may expand in other ways to be more universal platform. They're still more engineering, I think to be done. How has your role changed as the Kuban Eddie's project has scaled my role will I lead chronic for Cooper Netease, and it has been an extremely exciting ride. I would say I've enjoyed very much bringing this project in this product up and being there in the early days in bringing it up to kind of a more mature project where there are there's getting to be widespread usage. My role I would say in my team's role has shifted from guiding the initial technology working very closely with engineering to bring some of the innovation to market and explaining the innovation and changing mindsets, evangelizing, and kind of describing the benefits of containerization the benefits of this platform. You know, the everything we talked about here with regard to utilization and agility and portability. These were things that weren't obvious people were very rooted in VM's and two years ago. These things weren't very obvious in. So we worked on a lot of demos in use cases into to'real 's to teach users about this and also to learn about user environments to see how this technology could fit in. But I think in the last year. Year and a half that has shifted to actually use. There's no, and they come to us in, they tell us, you know, hey, I got high utilization. I'm being very productive like, okay, that's great. And what we're working on is, okay, how do we understand the enterprise requirements? How do we understand all of the security capabilities and build those capabilities and going back in and working with our engineering team to really prioritize and really understand the enterprise environment in which nutty is used? What are the other things that Kuban it is useful? What are the network constraints? What are the storage constraints? What are the traditional sort of hardware and legacy software that is is needs to interface with what are the things that we need to build to make it simpler for a new class of developers who, or maybe data scientists that are going to be using Cooper, nutty? How do we make it simpler for them to use the both at the towards the lower end of the stack, as well as at the top end of the stack, we've been innovating as a product team to get that under. Standing of the customer and build that roadmap for the product for Kuban Edison, and especially for GE to start to move into mainstream production regulated, highly secure enterprise environments, oftentimes in on premise use cases or in use cases that span enterprise and cloud an end, making it easier to use for a wet or class of developers. So those are the the ways in which my job has changed. And you know, I come from an enterprise software background, so I naturally gravitated towards, you know, making this software worked for large enterprises, providing a higher Soleil, increasing the security and reliability getting it into on premise environments and making sure that hybrid works well. But then we've also been innovating on the on the developer efficiency and the developer ease of use side of decay and Kuban Edison general. Now to some degree coober netties is you could just let Cooper Netease take take..

Kuban Eddie Kuban Edison Cooper Netease Kuban Cooper Maria DB developer GE two years
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

04:14 min | 2 years ago

"kuban eddie" Discussed on Software Engineering Daily

"And I think also tensor flow and airflow and red is, and so in it's based on sort of what's a good fit for coober Netease plus, what is their sort of immediate demand for in had dupe isn't yet, you know, it isn't a. Fit for Cooper, nutty. So I think that hasn't that is sort of at the level where we're integrating with it, but not really running it spark is and has been a very good fit. And so we have native support for spark and Kuban Eddie's is a native spit scheduler for spark. And then clustered databases obviously are natural fit or customer data stores are natural fit for Kuban Eddie's, but then also things like my sequel cockroach db and things that are Maria DB kind of more traditional kind of databases as well. And then of course, red is elastic search. These are things that people tend to use very commonly with what they're building on. Kuban is I think over time we may find that Kuban Eddie's expands to be a more universal platform that certainly come a long way already. It may expand in other ways to be more universal platform. They're still more engineering, I think to be done. How has your role changed as the Kuban Eddie's project has scaled my role will I lead chronic for Cooper Netease, and it has been an extremely exciting ride. I would say I've enjoyed very much bringing this project in this product up and being there in the early days in bringing it up to kind of a more mature project where there are there's getting to be widespread usage. My role I would say in my team's role has shifted from guiding the initial technology working very closely with engineering to bring some of the innovation to market and explaining the innovation and changing mindsets, evangelizing, and kind of describing the benefits of containerization the benefits of this platform. You know, the everything we talked about here with regard to utilization and agility and portability. These were things that weren't obvious people were very rooted in VM's and two years ago. These things weren't very obvious in. So we worked on a lot of demos in use cases into to'real 's to teach users about this and also to learn about user environments to see how this technology could fit in. But I think in the last year. Year and a half that has shifted to actually use. There's no, and they come to us in, they tell us, you know, hey, I got high utilization. I'm being very productive like, okay, that's great. And what we're working on is, okay, how do we understand the enterprise requirements? How do we understand all of the security capabilities and build those capabilities and going back in and working with our engineering team to really prioritize and really understand the enterprise environment in which nutty is used? What are the other things that Kuban it is useful? What are the network constraints? What are the storage constraints? What are the traditional sort of hardware and legacy software that is is needs to interface with what are the things that we need to build to make it simpler for a new class of developers who, or maybe data scientists that are going to be using Cooper, nutty? How do we make it simpler for them to use the both at the towards the lower end of the stack, as well as at the top end of the stack, we've been innovating as a product team to get that under. Standing of the customer and build that roadmap for the product for Kuban Edison, and especially for GE to start to move into mainstream production regulated, highly secure enterprise environments, oftentimes in on premise use cases or in use cases that span enterprise and cloud an end, making it easier to use for a wet or class of developers. So those are the the ways in which my job has changed. And you know, I come from an enterprise software background, so I naturally gravitated towards, you know, making this software worked for large enterprises, providing a higher Soleil, increasing the security and reliability getting it into on premise environments and making sure that hybrid works well. But then we've also been innovating on the on the developer efficiency and the developer ease of use side of decay and Kuban Edison general. Now to some degree coober netties is you could just let Cooper Netease take take..

Kuban Eddie Kuban Edison Cooper Netease Kuban Cooper Maria DB developer GE two years
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

02:06 min | 2 years ago

"kuban eddie" Discussed on Software Engineering Daily

"Often now and retail, and gaming and media it across all these industries. Every company has this type of team that starts to look at what's our next generation platform? How can we improve the productivity in utilization and get our business to the next level? And that's where they've typically bring in coober Netease. And then what they'll do is they'll set up a Kuban Eddie's environment. They'll of course tested out themselves. They'll start a few applications, and then they'll make it available to their constituents which could be other portions of the IT organization. Or it could be just developers in a particular business unit or across business units, and and then those developers will start to us. This typically, you know, it starts the way any other things starts. You know, I'll I'll use it for my web app. We're all use it for my front. Enter all use it for an ecommerce app, or maybe I'm writing something that needs a machine learning, you know, chat bought or something, or I'm I'm, I'm doing something that stateless that's typically wear. It will start and then they discover that actually can do more than that on this platform. Maybe I should try and run batch jobs, you know, that will actually be very efficient. I, you know, I, I'll only use what I need and seems like a good platform and then they'll find actually, yeah, there's a lot of documentation. There's a lot of history. There's a lot of tutorials. There's a lot of actually Zambales I can do that. Well, that works or if they're trying out something like server less or functions, this can be a great platform for that. And then eventually that platform team will get enough demand that they start to formalize the platform. They, of course, make sure behind the scenes that they get the security and compliance and networking capabilities. If they're building a hybrid cloud, they'll build a pathway to the public cloud so that you know this platform can be used in the public cloud and in in the on premise. And then they'll start to support more and more types of applications and kind of create an internal documents or set of documents about how to use the platform. And then I've seen many companies that have built an entire pass for their internal teams that's based on Cooper, nutty and for them to use Cooper Nettie. So I think the the thing that's different about Kuban Eddie's a relative to other technologies that I've seen is the rapid adoption of it..

Cooper Nettie Kuban Eddie
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

01:31 min | 2 years ago

"kuban eddie" Discussed on Software Engineering Daily

"As your container service simplifies the deployment management and operations of coober netease eliminate the complicated planning and deployment of fully orchestrated containerized applications with cooper netease you can quickly provision clusters to be up and running in no time while simplifying your monitoring and cluster management through auto upgrades and a built in operations console avoid being locked into any one vendor or resource you can continue to work with the tools that you already know so just helm and move applications to any coober netease deployment integrate with your choice of container registry including azure container registry also quickly and efficiently scale to maximize your resource utilization without having to take your applications offline isolate your application from infrastructure failures and transparently scale the underlying infrastructure to meet growing demands all while increasing the security reliability and availability of critical business workloads with azure to learn more about azure container service and other azure services as well as receive a free book by brendon burns goto aka dot ms slash s e daily brendon burns is the creator of coober netease and his e book is about some of the distributed systems design lessons that he has learned building kuban eddie's that e book is available at aka dot ms slash s daily.

brendon burns kuban eddie
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

01:31 min | 2 years ago

"kuban eddie" Discussed on Software Engineering Daily

"As your container service simplifies the deployment management and operations of coober netease eliminate the complicated planning and deployment of fully orchestrated containerized applications with cooper netease you can quickly provision clusters to be up and running in no time while simplifying your monitoring and cluster management through auto upgrades and a built in operations console avoid being locked into any one vendor or resource you can continue to work with the tools that you already know so just helm and move applications to any coober netease deployment integrate with your choice of container registry including azure container registry also quickly and efficiently scale to maximize your resource utilization without having to take your applications offline isolate your application from infrastructure failures and transparently scale the underlying infrastructure to meet growing demands all while increasing the security reliability and availability of critical business workloads with azure to learn more about azure container service and other azure services as well as receive a free book by brendon burns goto aka dot ms slash s e daily brendon burns is the creator of coober netease and his e book is about some of the distributed systems design lessons that he has learned building kuban eddie's that e book is available at aka dot ms slash s daily.

brendon burns kuban eddie
"kuban eddie" Discussed on .NET Rocks!

.NET Rocks!

01:39 min | 3 years ago

"kuban eddie" Discussed on .NET Rocks!

"Oh i think join so there's gabe in then there's brennan burns who worked on kuban eddie's they're actually a part of the core right brennan burns open source guy but he came to work for microsoft yes he's the director and gave runs the the team okay waking on ks and so we have some of the core founders of the of the projects actually working here which is really cool we learn what is helm helmet's away to use a template to basically declared a wli install and get to a desired state of what's in your clusters let's say i want to have a side car running a particular logging agent i would declare that in helm it's almost like it's almost like a boot or package manager but for a cluster right it's the manifest that describes how the cluster is going to behave you can figure ports you can figure what is publicly exposed versus privately exposed so you know if you want your agent assessed right and it's also becoming kind of unit of extensively in sharing so you see that a lot of common to policies or rigs or just app type so now described as helm charts and you can go recreate an entire distributed application using a home china i think it's worth mentioning i can't remember we said it or not so politics from repeating it but it's a templeton language for then generating the committee's manifest so you can author your kuban eddie's manifests right and then i could check that in and now when you want to change three things when you go.

gabe kuban eddie ks brennan microsoft
"kuban eddie" Discussed on .NET Rocks!

.NET Rocks!

01:39 min | 3 years ago

"kuban eddie" Discussed on .NET Rocks!

"Oh i think join so there's gabe in then there's brennan burns who worked on kuban eddie's they're actually a part of the core right brennan burns open source guy but he came to work for microsoft yes he's the director and gave runs the the team okay waking on ks and so we have some of the core founders of the of the projects actually working here which is really cool we learn what is helm helmet's away to use a template to basically declared a wli install and get to a desired state of what's in your clusters let's say i want to have a side car running a particular logging agent i would declare that in helm it's almost like it's almost like a boot or package manager but for a cluster right it's the manifest that describes how the cluster is going to behave you can figure ports you can figure what is publicly exposed versus privately exposed so you know if you want your agent assessed right and it's also becoming kind of unit of extensively in sharing so you see that a lot of common to policies or rigs or just app type so now described as helm charts and you can go recreate an entire distributed application using a home china i think it's worth mentioning i can't remember we said it or not so politics from repeating it but it's a templeton language for then generating the committee's manifest so you can author your kuban eddie's manifests right and then i could check that in and now when you want to change three things when you go.

gabe kuban eddie ks brennan microsoft
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

01:31 min | 3 years ago

"kuban eddie" Discussed on Software Engineering Daily

"Money depending on kuban eddie's and so they take that work very seriously ccf's role there is just a facilitator it i will mentioned some other work that we've done that's a little newer where to cnc f projects coordinator and on voi requested that we fund us third party security audits and so we used a company cure fifty three and they went off and coordinates they found a serious bug that they fix but more generally they had some other suggestions and fixes all those were taken into account and the the reports are public so you can you can link to them in the show notes but that's some work that we're we're interested in doing for more of our projects i mean security audits they're only ev ever a point in time it's still humans doing it so there's nothing perfect about it but we're really thrilled to be able to contribute some of the resources of our members back into this projects and to help them along and i will say that overall the summary reports were extreme namely positive that they spoke very highly in terms of the quality of the code and the conscientiousness of the design and and did recommend them we start off talking about service that's one of the working i believe it's a working group right yeah it's it's a working group within the cnc f and another working group that i attended a couple of talks on was storage storage is interesting because well maybe you could just tell me why why you believe storage is storage on coober netease.

kuban eddie ccf coordinator
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

01:31 min | 3 years ago

"kuban eddie" Discussed on Software Engineering Daily

"Money depending on kuban eddie's and so they take that work very seriously ccf's role there is just a facilitator it i will mentioned some other work that we've done that's a little newer where to cnc f projects coordinator and on voi requested that we fund us third party security audits and so we used a company cure fifty three and they went off and coordinates they found a serious bug that they fix but more generally they had some other suggestions and fixes all those were taken into account and the the reports are public so you can you can link to them in the show notes but that's some work that we're we're interested in doing for more of our projects i mean security audits they're only ev ever a point in time it's still humans doing it so there's nothing perfect about it but we're really thrilled to be able to contribute some of the resources of our members back into this projects and to help them along and i will say that overall the summary reports were extreme namely positive that they spoke very highly in terms of the quality of the code and the conscientiousness of the design and and did recommend them we start off talking about service that's one of the working i believe it's a working group right yeah it's it's a working group within the cnc f and another working group that i attended a couple of talks on was storage storage is interesting because well maybe you could just tell me why why you believe storage is storage on coober netease.

kuban eddie ccf coordinator
"kuban eddie" Discussed on IOT Podcast

IOT Podcast

01:31 min | 3 years ago

"kuban eddie" Discussed on IOT Podcast

"Okay that's how that works and so in effect what that does is it means i can now have a communities coster spans cloud and edged and deploy workloads across both at the same time see so then coober netease actually becomes your os for the cloud in an orchestrator for the cloud on the edge okay so this is very exciting in this gets into this concept that i've been involving with analysts and smart people like yourself here which is we're going for the iot world we're going to have the cloud you're gonna have the not chat big edge fat edge big edge and then you're gonna have your silently devices not devices and what you found with the virtual coup blitz is a way to kind of fan that all with to wrap it all in this in our nomenclature the way we think about it is you know you have cloud you have edge which can run these containers from the cloud and then you have iot devices which are really purpose ville connected to sensors and things like that so we see the kuban eddie's clusters spanning cloud in but then being able to that entire cluster being able to work with iot device okay so it doesn't actually correct run out in it correct is typically iot little iot devices don't run containers like i said there are light bulbs now that run docker there are some of those yet so okay i came here hoping to hear newer use cases and i feel like you and i have talked in the past about rockwell in her she's sure in there is a continuing issue in the business face where rest pilot the heck out of this and then nothing ever goes very far.

kuban eddie rockwell
"kuban eddie" Discussed on The Internet of Things Podcast - Stacey On IoT

The Internet of Things Podcast - Stacey On IoT

01:31 min | 3 years ago

"kuban eddie" Discussed on The Internet of Things Podcast - Stacey On IoT

"Okay that's how that works and so in effect what that does is it means i can now have a communities coster spans cloud and edged and deploy workloads across both at the same time see so then coober netease actually becomes your os for the cloud in an orchestrator for the cloud on the edge okay so this is very exciting in this gets into this concept that i've been involving with analysts and smart people like yourself here which is we're going for the iot world we're going to have the cloud you're gonna have the not chat big edge fat edge big edge and then you're gonna have your silently devices not devices and what you found with the virtual coup blitz is a way to kind of fan that all with to wrap it all in this in our nomenclature the way we think about it is you know you have cloud you have edge which can run these containers from the cloud and then you have iot devices which are really purpose ville connected to sensors and things like that so we see the kuban eddie's clusters spanning cloud in but then being able to that entire cluster being able to work with iot device okay so it doesn't actually correct run out in it correct is typically iot little iot devices don't run containers like i said there are light bulbs now that run docker there are some of those yet so okay i came here hoping to hear newer use cases and i feel like you and i have talked in the past about rockwell in her she's sure in there is a continuing issue in the business face where rest pilot the heck out of this and then nothing ever goes very far.

kuban eddie rockwell
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

01:42 min | 3 years ago

"kuban eddie" Discussed on Software Engineering Daily

"Okay so we need a registry in the u so it's super easy to run one but then what do we use for the day store we can use gcp but then we earth's our s but then we would need to keep all of our images in sync so we ended up doing was we built this i guess it's like a proxy and what you do is you point things that use s three to this proxy and what it does is i checks is this file s if it is return on immediately if it's not pull it in from s three and while you're copying it over saving g c s so we've got this registry in kuban eddie's on the eu that's pulling the images from s three storing g c s so that we can use it later the one rankle was sore we're using that we're running that proxy in docker in communities like we can't host it you know in our own things so we ended up having a polish that one and have that'd be public but kind of cool does that answer your question those so does brooke extension inside coober daddy's will that make a difference for you because now being able to keep this file systems within kuban eddie's that's a great question we haven't really looked at any providers other than just the normal ubs in aws and the pd one gcp probably down the road we will be looking at gloucester ourself or or one of those things are using aws lambda for anything we're using it a little bit for things related to our content platform i actually don't have a lot of insight into it other than we're using it.

kuban eddie eu aws kuban
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

01:42 min | 3 years ago

"kuban eddie" Discussed on Software Engineering Daily

"Okay so we need a registry in the u so it's super easy to run one but then what do we use for the day store we can use gcp but then we earth's our s but then we would need to keep all of our images in sync so we ended up doing was we built this i guess it's like a proxy and what you do is you point things that use s three to this proxy and what it does is i checks is this file s if it is return on immediately if it's not pull it in from s three and while you're copying it over saving g c s so we've got this registry in kuban eddie's on the eu that's pulling the images from s three storing g c s so that we can use it later the one rankle was sore we're using that we're running that proxy in docker in communities like we can't host it you know in our own things so we ended up having a polish that one and have that'd be public but kind of cool does that answer your question those so does brooke extension inside coober daddy's will that make a difference for you because now being able to keep this file systems within kuban eddie's that's a great question we haven't really looked at any providers other than just the normal ubs in aws and the pd one gcp probably down the road we will be looking at gloucester ourself or or one of those things are using aws lambda for anything we're using it a little bit for things related to our content platform i actually don't have a lot of insight into it other than we're using it.

kuban eddie eu aws kuban
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

02:02 min | 3 years ago

"kuban eddie" Discussed on Software Engineering Daily

"Packed with so you can stand up an api server just right next to the cosmo's db setup that you have and then on google maybe you're using google's i dunno pub sub service together with your cougar netease instance how do you do you anticipate these different coober netties clusters on different cloud providers and how do you see those different kuban as clusters communicating with one another it's actually really interesting because i do see that happening lot both in people experimenting some people running multi cloud because they have a cuban cluster running in their own data center then you know they're leveraging cloud provider for some other service but you know things like coober nineties it's becoming a commodity right you know all the major cloud providers have a coup vanities offering you know we're bound to see more and more companies kind of spin up you know kind of in the same vein as you know when slices and dm's became popular right like lots and lots of people offered those things so i think we're going to continue to see more people off for kuban eddie's in and it's going to be a commodity and that's a good thing but like you said some of the power that comes in cloud providers is the other services they offer you know they've got a lot of machine learning things in cognitive services and databases that you don't have to worry about operating that can run just massive scales and their offerings just continue to grow you know we're going to get into doing things like quantum computing and things like that and these are all things you'd want to mess with locally so i do see that people are probably going to do that and you know a lot of the things that differentiate cloud providers are those additional offerings so i think as all of this offerings kind of continue to build out people will end up having clusters in each i would think you'd probably try to isolate these things in the same way that you do microsoft is where they each kind of have their own role and there's kind of light communication between them it's possible people end up using federation federation is still going kind of.

google kuban eddie microsoft
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

02:26 min | 3 years ago

"kuban eddie" Discussed on Software Engineering Daily

"As your container service simplifies the deployment management and operations of cooper netties eliminate the complicated planning and deployment of fully orchestrated containerised applications with cooper netties you can quickly provision clusters to be up and running in no time while simplifying your monitoring and cluster management through auto upgrades and a builtin operations consul avoid being locked into any one vendor or resource you can continue to work with the tools that you already know such as helm and move applications to any kuban eddie's deployment integrate with your choice of container registry including asuar container registry also quickly and efficiently scale to maximize your resource utilization without having to take your applications off line isolate your application from infrastructure failures and transparently scale the underlying infrastructure to meet growing demands all while increasing the security reliability and availability of critic gold business workloads with azure check out the azure container service at a k a dot ms slash acs that's aka a dot m s slash acs and the link is in the show notes thank you to azure container service for being a sponsor of software engineering daily same goads izzy the architect at box and was one of the founders of the company sand welcome to software engineering daily thanks very much things rowden i started a high level of how things work at box and then we'll get into a discussion of cougar netties so i think most people know what box is it's a files storage and management service most people blessing it probably used it go as deep as you want but what happens when i upload a file to box when i am just a user and i upload a file what's happening on the backend sure so at a high level our companies started before public cloud back into hasn't five so we run and managed most for infrastructure we have an migrate some stuff the public loudly interview s and a little bit on gcp editor by predominantly we run on infrastructure so you know for the most part when someone opens a file it's going to reach one of our data center.

kuban eddie editor cooper
"kuban eddie" Discussed on Software Engineering Daily

Software Engineering Daily

02:41 min | 3 years ago

"kuban eddie" Discussed on Software Engineering Daily

"Yes saw amazon on those will list for a while because people have made investments in them are going to back away overnight as your container service simplifies the deployment management and operations of cooper netties eliminate the complicated planning and deployment of fully orchestrated containerised applications with cooper netties you can quickly provision clusters to be up and running in no time while simplifying your monitoring and cluster management through auto upgrades and a builtin operations consul avoid being locked into any one vendor or resource you can continue to work with the tools that you already know such as helm and move applications to any kuban eddie's deployment integrate with your choice of container registry including asuar container registry also quickly and efficiently scale to maximize your resource utilization without having to take your applications off line isolate your application from infrastructure failures and transparently scale the underlying infrastructure to meet growing demands all while increasing the security rely bility and availability of critical business workloads with azure check out the azure container service at a k a dot ms slash acs that's aka dot m s slash acs and the link is in the show notes thank you to azure container service for being a sponsor of software engineering daily so you've mentioned with the opensource aspect of dealing with dais that you you learned a lot you know this is something i'm kind of dealing with we have this set of opensource projects with software engineering daily and there are some things that i have found to be unexpected even with just this project which is fairly simple it's like these apps that people can open ten people can download and basically it's opensource ways of running software engineering daily apps on their phones and they're just simple things like always keeping the repositories in a state where people can stop by in check it out and you know things like going through the get hub issues in a timely manner but this is even with just a simple projects i imagine with dais there's a whole lot more volume of the stuff that you had to deal with an also you were building a company around the system so what were the pros and cons to building a company.

amazon kuban eddie cooper