18 Burst results for "Coober Netties"

"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

04:10 min | 1 year ago

"coober netties" Discussed on Software Engineering Daily

"I really don't have any problem. Promoting g too. I because i have been a power user of their services and i gotta tell you it will accelerate your development cadence. Were talking at a pretty high level here. And i wonder from for from somebody who's been watching distributed systems since they were coming into their own. How much has coober netties helped with the development process of this kind of stuff that we're talking about yes so first of all. I'm going to. I hope you tell me. It's over hyped. Is it over hived. I'm not a coup brunetti fan boy in the sense that i think it's inelegant in a lot of ways but it's one like there isn't a reasonable alternative that you can tell people this is this is this is the heterodox infrastructure podcast today. So i've made my peace with cooper netease and in fact our cloud services are built on cooper netease with the eye towards if somebody wants to run astra cassandra or astra pulsar on their own. Vm ware or red hat who bonetti's installation then we can give them The operators control plane to do that. So if kuban as didn't win because of superior technology wide kuban as win. It's backed by google and this critical mass past which you're kind of spitting into the wind right and i think we hit that point to a couple of years ago but wasn't it. At least the least bad architecture was mazo significantly. Better was cloud foundry better. You could be right. I'm not looking to argue. The details of how good the architecture is all right fair enough. But i'm sorry. So i got a little sidetracked there but does it help to at least have a standardization. You at least we have. Maybe the lennox distribution we built socks. But at least we're all using it. I'd have to think about that. Analogy to see how much i agree with it. But yes it does help a great deal to have that common layer but you can start building against okay so you sell people pulsar. Do they care or has costco. One net spitting into the wind thing. So yes and yes and no i think in that order. So yes we're selling pulsar. Two people in two ways were selling cloud service that i mentioned and we're also selling a luna streaming support in distribution for people who wanna run pulsar themselves and so what people don't like is rewriting code right so if i were going there and saying hey i want you to rewrite everything these pollstar. I think that would be super difficult. Despite the technical advantages that pulchra has architecturally. But that's not what i'm telling people i'm telling people. Hey no i have a james implementation that you can start using. Tomorrow i'm saying i have a pulsar. Cluster that speaks the kafka api and so until my messages. You don't have to rewrite your code. You can move it to pulsar without doing that. Rewrite and so yeah. People are very interested in that. What's the process of getting to a point where you feel comfortable handing this over to people as critical infrastructure. So i'll back up a little bit to the one of the first things. I said here where i said. That pulsar is bulletproof. And i said that an asterisk next to that so the asterisk is that pulsar is still at the kind of adolescent stage of open source projects where it's adding new features faster than it can keep those stable and what i mean by that is so i mentioned we have a distribution of pulsar luna streaming the version of luna streaming. That's outright now is based on pollstar. Two dot six dot two were working on a version of luna streaming based on pulsar. Two seven two. That's going to be out. In the next couple of weeks. Meanwhile apache pulsar released version. Two dot eight.

cooper netease kuban astra Vm ware bonetti mazo brunetti pulchra costco google
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

04:09 min | 1 year ago

"coober netties" Discussed on Software Engineering Daily

"I really don't have any problem. Promoting g too. I because i have been a power user of their services and i gotta tell you it will accelerate your development cadence. Were talking at a pretty high level here. And i wonder from for from somebody who's been watching distributed systems since they were coming into their own. How much has coober netties helped with the development process of this kind of stuff that we're talking about yes so first of all. I'm going to. I hope you tell me. It's over hyped. Is it over hived. I'm not a coup brunetti fan boy in the sense that i think it's inelegant in a lot of ways but it's one like there isn't a reasonable alternative that you can tell people this is this is this. Is the hetero infrastructure podcast today. So i've made my peace with cooper netease and in fact our cloud services are built on cooper netease with the eye towards if somebody wants to run astra cassandra or astra pulsar on their own. Vm ware or red hat who bonetti's installation then we can give them The operators control plane to do that. So if kuban as didn't win because of superior technology wide kuban as win well. It's backed by google and this critical mass past which you're kind of spitting into the wind right. And i think we hit that point to a couple of years ago but wasn't it. At least the least bad architecture was mazo significantly. Better was cloud foundry better. You could be right. I'm not looking to argue. The details of how good the architecture is all right fair enough. But i'm sorry. So i got a little sidetracked there but does it help to at least have a standardization. You at least we have. Maybe the lennox distribution we built socks. But at least we're all using it. I have to think about that. Analogy to see how much i agree with it but yes it does help a great deal to have that common layer but you can start building against okay so you sell people pulsar. Do they care or has costco. One net spitting into the wind thing. So yes and yes and no i think in that order. So yes we're selling pulsar. Two people in two ways were selling cloud service that i mentioned and we're also selling a luna streaming support in distribution for people who wanna run pulsar themselves and so what people don't like is rewriting code so if i were going there and saying hey i want you to rewrite everything these pollstar. I think that would be super difficult. Despite the technical advantages that pulchra has architecturally. But that's not what i'm telling people i'm telling people. Hey i have a james implementation that you can start using. Tomorrow i'm saying i have a pulsar. Cluster that speaks the kafka api and so until my messages. You don't have to rewrite your code. You can move it to pulsar without doing that. Rewrite and so yeah. People are very interested in that. What's the process of getting to a point where you feel comfortable handing this over to people as critical infrastructure. So i'll back up a little bit to the one of the first things. I said here where i said. That pulsar is bulletproof. And i said that an asterisk next to that so the asterisk is that pulsar is still at the kind of adolescent stage of open source projects where it's adding new features faster than it can keep those stable and what i mean by that is so i mentioned we have a distribution of pulsar luna streaming the version of luna streaming. That's outright now is based on pollstar. Two dot six dot two were working on a version of luna streaming based on pulsar. Two seven two. That's going to be out. In the next couple of weeks. Meanwhile apache pulsar released version..

cooper netease kuban astra Vm ware bonetti mazo brunetti pulchra costco google
"coober netties" Discussed on Back From The Future Podcast

Back From The Future Podcast

03:47 min | 2 years ago

"coober netties" Discussed on Back From The Future Podcast

"Some of them are actually hands on keyboard. They are interactive and you have to simulate what you'd be doing in the real world, so that's also a plug for certifications. Okay, so you've convinced me. It's time to go into the tech field, but I WANNA make some money here. What? What are some really high paying career paths in the tech field right now you mentioned the word before Dj at his will, but right now the trendy fields with a high paying fields. It's cloud. It's cloud right now. We are living in a cloud world in the cloud is just someone. Else's computer is just a computer that's not in. You're not in your data center. It's in someone else's. That's basically what the cloud is, but because the different clouds. Clouds used their own technology their own terms their own way of doing things. If you could speak the language, you can get paid so if I was started off right now. I would really take a look at some of the cloud jobs like you can't go wrong like before with Amazon aws or Microsoft Azure. If you want to be someone who can be general, you know, try to look at the engineering science. You're the guy who can stand. Stand up new service data workstations in the cloud. Now you can make pretty good money starting off like you know six figures easily, but again it depends on where you are in your career. What your resume looks like says even if you were to be brand new up the gate. If you just have it on paper, you don't have the experience you know you. You can get paid less so those do. Go Hand in hand you can also. Also, of course, be a programmer. If you are a python programmer, Java Nojeim Gas android still IOS. You can make a whole lot of money were always looking for more deb, so if you've got that accurate look into development, if you like it operations any like program, look at something like Coober netties or devops, you can combine multi discipline and get lots of work also right now. Dj Data Scientists. It's hot big big data. It's hot so those. Those, are you know six fields I would definitely look into right now if you WANNA know who's paying the most right now like I, said before it, a couple of podcasts before like I'm a big fan of Wayne Gretzky. You WanNa Skate to where the puck is going, so go online to your research. See what is an emerging tech and see where you think the tech might go in the next couple of years like right now I know robotics. Robotics is huge as well and that's it also people who work on drones. That's it they get paid so one of the things that I felt like you may have touched on very lightly. Was information security that field are filled I say also very high paying and the fact that if you're a developer, you're typically not tied down to an office. A lot of these jobs can be remote, so you can live where you would like to. To live in earn very high pay potentially in a low cost of living area now. Can you get into tech related career path with no training or experience? Can you give into some of the fields that run right alongside these career paths yet? I think you can I think that anyone can can learn anything if they are not motivated..

programmer Wayne Gretzky Dj deb Amazon Microsoft developer
Go Networking with Sneha Inguva

Software Engineering Daily

09:41 min | 2 years ago

Go Networking with Sneha Inguva

"Guba. Welcome to software engineering daily. Thank you thank you for having me here have been a huge fan for a while so. I'm super excited and humbled to be on the show right. Well happy to have you on you work at Digital Ocean which is a cloud provider. Give me a few examples of engineering problems that you've worked on so digital ocean. We are cloud hosting provider. We have a variety of products in different areas for example with storage with networking as well as compute. Which is probably. I guess what most people are familiar with who used digital ocean we have droplets serve virtual machines that they can use but the interesting thing as cloud hosting providers that it's a little different from other companies in which in that we have both physical hardware issues we also have software issues and then we also have a web application so we've had interesting problems kind of all over the place when I joined the company. I wasn't actually network engineer. I was working on. One of the internal delivery. Teens is what we call the and on that team the biggest problem we were addressing was the difficulty in deploying and updating applications so namely working with Kubrick so that was definitely an interesting problem because I think we addressed. Both you know the challenge in building an abstraction layer on top of Kuban as that increased the just ease of deploying because before that people use chef chef was a little complicated in general and then on top of that also getting buy in from different teams to kind of use this new internal tool that we had so that would. That's kind of one of the problems we've had that we've addressed as you've mentioned digital ocean is built around these abstractions called droplets. Can you say much about what droplet is? Is it a VM? Is it a container? What am I actually interfacing with? When I spin up a digital ocean instance of course so it is a virtual machine. I think droplet just our marketing speak for everything oceanic themed in our company but it is essentially a virtual machine that is I guess. Technically Co located on servers with other virtual machines and you can spin up really in any location around the world. I think we have about thirteen data centers. So that super fine I I also heard you mentioned container so right now. We don't have containers as a service but we have coober. Netties is a service so technically speaking you could manage your containers as well although droplet itself is just a ritual machine. Got It now when you join a company. It's always tricky to find the bounds of what you should learn. And what you should know. R- it's hard to know just how deep to go and I know that when one of these virtual machines spun up. There's a ton of stuff that is going on under the hood. What was your process for figuring out what to learn the the life cycle of a user spinning up a VM. That's a really good point. In fact I think I think we still do this. When someone we have a for networking at least we have a really good on boarding process. Or when I joined the company not a networking we also had still had a pretty good on boarding process but it was more generic and there is in fact I guess. An on boarding session called how. The cloud works where an engineer who's been at the company for a while actually goes through the entire process and kind of goes through all the micro services that I guess receive a request and send a response. You know down to the schedulers that actually are scheduling the droplet placement on a particular hyper visor. Down to everything. So the thing is I think most people probably have a general idea of the different services that are being touched but then when it comes down to the nitty gritty of how exactly he's Networking Setup Hauser. Sdn configured all of that. I don't think unless you're on that specific team. You are aware so. It's it's kind of a t shaped process in a way so you have a general like breath of knowledge of how I guess the cloud works quote unquote but when it comes to the nitty gritty details. You probably have a very good idea of just your specific area. And I think it's impossible to have a very deep knowledge of absolutely every single service when you're at a company this large with this many micro services and with this many domains of expertise totally now. The reason I want to have you on the show is because I saw some talks. The you gave one specific talk about networking and the term networking can mean a lot of different things. But I know that now working at a cloud provider and you being a systems engineer working at a cloud provider. You probably have some insights on the engineering that goes into the actual nitty gritty of something spending up within digital ocean. What does networking mean at a cloud provider? What does that term networking so networking at a cloud provider? I think has two layers. There's of course the physical infrastructure that is set up so of course I think every cloud provider has physical switches physical address physical gateway so that is definitely one layer but then another thing that you have to consider especially at a cloud provider where you are dynamically. Creating and deleting virtual machines is that you are constantly adding different paths for networking packets traverse and removing them as well. So that's where software defined networking comes in and that's a completely different layer that you have to consider especially at a cloud provider and in fact at digital ocean. We actually have a team that deals with a lot of the physical details when it comes to physical switches in our data centers but we also have a SDN team which has a lot of sub teams that deal with a lot of the micro services that are interfacing and communicating with obvious open switch which is our virtual switch of choice that are actually making a lot of our networking products. Possible such as you know such as VP see or firewalls or even DHCP. A lot of these different things about some of the lower level networking concepts that you needed to know to build some of the projects that you've built within digital ocean. Of course so I'll just take you through. I guess when I first joined the networking team we were coming out with a product called. Bring Your own image so previously when people typically spin up a virtual machine or droplet they can select predefined image whether it's a boon to or I don't think we have Microsoft but a different version of a boon to or one of many different options however with Byu Hawaii we started giving them the option of bringing their own image. So the only issue with that is when we control the image ourselves. We can kind of control the cloud configuration meaning allocating IP addresses and setting up a lot of configuration. But when they're bringing their own image we need a way to dynamically allocate Ip addresses for those droplets using that image and that's where the DHCP protocol came in. And that was something that I had heard of. But I wasn't super familiar with but in general I guess whenever you're building a new networking product. That's using a new protocol my first step typically is to read the RMC so I pulled up the DHCP are of C and then the DHCP C. Which is a little different and started to learn about the protocol and I guess most people at home are probably familiar with it when they log into their computer and they fire up the Internet their ISP Rod are actually allocates Ip address for their home computer and so that's essentially using the DHCP protocol so we were implementing our own. I guess a hyper visor level demon to do that for different droplets in our data center and so that was something that I started to learn about. And then the other thing. When you're a cloud hosting provider is you start to learn about perhaps the ways in which you might have abusive actors and kind of look into security and so that was very interesting and then you you start to do a lot of load testing and try to figure out how to mitigate any possible issues so that was also something else. I started to look into when it came to the server that phrase you mentioned. Rfc reading the RFC. I've read some core answers and wikipedia recommendations about you WANNA learn networking concepts you should read the RFC which chance for the request for comments. Why is that the best path to learning about networking protocols? I mean that is fundamentally where the networking protocols were designed and some of these protocols redesigned like decades ago so I think that of course you could read wikipedia articles encyclopedia articles youtube videos. All of those are helpful. But I think that going to kind of the original source of where this communication protocol was defined. And of course to be honest the first time you read through any networking RFC. It won't one hundred percent make sense so obviously going through it marking up everything you don't understand which then and then of course every RC is somehow linked Tillich twenty other RC's so then go jumping to another RFC to kind of understand. Maybe another protocol that is used within a particular protocol kind of helps you build sorta like in a mental map or like a mental knowledge tree of what that protocol actually does what it

Digital Ocean Engineer Kuban CO Kubrick Netties Systems Engineer Tillich VP Microsoft
"coober netties" Discussed on Data Engineering Podcast

Data Engineering Podcast

13:22 min | 3 years ago

"coober netties" Discussed on Data Engineering Podcast

"May 'em for example on strings we have a very powerful kodak a cold a low carb analogy and what this does is it basically if you let's say you have as a string value for region in some data set while there's a finite number of geographic regions in in most cases say you know say maybe you have five hundred if these well if you apply this low kurt analogy a data typer we call it look harden like data type but it's really a transformation you play this column what it will do is automatically convert those region values to injure and maintain look up table for you that did it automatically been associated so that when you do a select you see the region name but underneath it's just it's just comparing and a selecting using managers so it's basically three levels you know you're starting out you just make you could make a basic choice about a partition key in short order but but just have a rough estimate on they you know on the on the data types you could never find the data types and then you could apply different codex and other transformations on the data and that's that's kind of the way that you could get to really efficient scheme i i think that it's there's also another question in modeling that comes up very commonly in many cases and that's what to do when you have variable schemer and that's something that alexander's sexually worked on quite a bit and alexander i think that's one where you can offer some interesting comments yes i think so i in modern day to call a some users say used to a kim a less data poisonous are little boy this week's announcement came alive cosigned for example and a or some tendency to think of ways it's like influence or anything else and they comfortably call us so if he's grown to wonder if they can do the same and in most cases makes sense because a forgotten will consider them a monitor including a lot of different devices oldie license making their own parker is insulin and it's been hot micro scheme other works for case or it would have a phone call on the table so instead of doing house is just to update check in the race so you might all right well he's earned it rains are ordering multiple arrests were decent size is kind of gets away gives you away from the column i stopped smoking calls because all right is solved altogether a block a dot com from these and a walk wish scheme a less a daytime and sort of seen as q better players and if you sign up for months at las vegas you can then mike channelize song a season where do you see the interviews from those humidity race as a simply columns kinda donna off doors kinda done associate with no downtime john's by off the table and costs will start with police in those into this the important for example where you end up in fact a quaint in the eighties and those symptoms much more fun than like a so it's just kind of hybrid approach right so you start to leave this the right a or enraged and probably some other a common represents like time stands a tumor and then you talk such and such and data 'em for the former similar to what percent order skimmers things and then senior dating apps quote in so long in house allows it yeah we have users using exactly this approach for example the process logs a you know keep the date in a raise and then use this materialized column feature that's a actually another really killer click house feature of the fact that you can you can do it with zero downtime and basically you know sort of instantly began to to refer to to the new column which correct type is is a great feature yeah and all of these capabilities as far as the flexible modeling options puts me in mind of the current trend as far are as the shift away from e t l n t e l t where you land the raw data into some source where in a lot of cases it's a data lake but with some of these modern cloud capable data warehouses such as snowflake db i know encourages you to just load the raw data into it and then do the transformation after the fact and so it seems that courthouses another case where you're just load the raw data into some source table and then take advantage of these scheme evolution capabilities are given that elation capabilities abilities for then being able to transform the data into a more process foreign per being able to use for more advanced analytics yeah that's why i'm saying that he's just from a use case what it calls a bucket hollows kind of a it's a it's it's a rose oh yeah and i think we're definitely not i mean there are systems out there which will actually allow you to do you know you'll have met a store which all you know sort of implement scheme up on reid so that you may have just text files that this is the this is the hindu i've model that sort of the far end of the spectrum and cookouts is absolutely not designed or not intended the follow that pattern we do require forget highperformance you really going to have to put the data into into click house storage format you know we kind of course referred a data in other databases but those tend to be smaller tables so we do expect that people are going to eventually load most of their data into click house since and stored in the native merge tree format so we're kind of closer to a in that sense really closer to eighty traditional database then then for example you know some of the things is some of the systems are designed jewelry data directly office three and then as far as getting quick has deployed i know you said in terms of just be initial on boarding you can do not get installed but for actually putting it into a production contacts i'm wondering what's involved in some it'd be edge cases that users should be aware of and some of the tuning capabilities that would be useful to point out a question so try to rough house a boston six seconds and the yukon financial you're welcome for or higher and silver calls words around portable end when go into production those considerations so it's not always a whole nothing i should not always the best one and helps looking house which is quite often it's it's more like feasability for but she didn't start construction set up for auction for months policies but up too much for hanging in there muddled shores of course in a hang onto some tone optimizations july so a ship your cash to for instance like a date is you're well and so sometimes if she is a certain those insurance this is not the case but so you're kind of go with calls to clouds as well including public wealth like amazon or google cloud uncut college or sleepnumber there well she are again we provide ways other groups storage a social and it's something that not the phones as fast as unbearable uncomfortable bare metal show horse i'm gonna sutton expense organization but it's it's very convenient flight so just deployed to the cloud turkish evidence now santa claus to extradite don't have a reason to go somewhere else what do we see now and the new moms out there is a growing interest to a decline could call us in they're not just because lawsuit to run this same college country gration everywhere including quebec not just some dental the cuban that just in quotes in public in france laws and a wow what could also be exactly the same environment where she could columnist at the new more and more plus loan but they're not this mission to the moon is focused on the data house or i will buy stuff comes around disingenuous toll and a multiple reasons for that reason is that a potential vacations and occasional stone at your place in utica but not this and the house was also a ceo women boston in not just the conflict in you know then kuwait tobias so a kind of silicon but it's it's going to cost share a lawsuit from everywhere ends up too much but you chore use case scenario is shipped muscle formers your goal wildly you shift mice will continue to go to another way and the title sonata huge so quebec disorder is not significantly contacted their mental illness and was on a song yeah i think the performance of clicks on coober netties it's been very enlightening southern coober netties operator that we've developed takes what is naturally a very complex deployment which is in anything particular decree quick house it's just the deploying data you applications were state on on coober netties involves creating a lot of resources and managing them so so a the coober netties operator that we've developed reduces these to a single custom resource it's easy to configure and a finger deploy an update but what we've definitely seen is is that well first of all these gives you huge flexibility because you could quickly spin up clusters you concrete clusters in less than a minute depending on their size of course but you also have the we've also seen pretty promising performance results that they for example when we run test that run on coober netties versus a running on vm so coober netties on amazon bm's versus just running on straight pm's which is virtually no difference in the in the performance so that's actually something that looked very promising an we we do 'em we are working with customers who are having major deployments on cooper netease and we definitely see this is a growing a as a growing a platform for data warehouse in future and and as far as we sort of use cases and instances of quick house and customers they you've worked with i'm wondering what have been some of the most interesting or unexpected innovative ways you've seen it used i'll give you one that just sort of blew my mind about four days after i got on the job a we have a customer that does a they developed custom fraud detection systems for banks and credit card processors so the way they used click house is they bring it in and the first thing they'll have to do is in order to build the system in select appropriate models and things like that for for doing fraud detection they actually need understand what data that financial institution available so they'll just ask them to give him a few csv files containing more less complete samples of data they'll just throw it in the quick house and start playing with it and looking around and this is exactly what i'm talking about when i you know when i i was discussing how straightforward modeling conveyed it you could literally just started strings this is this is exactly what they do they just throw the cs vn they don't start playing around with it i have not ever seen a database they can do this on this scale so it's a really interesting in an innovative way they used yes a sort of a good example and some i saw i saw a record in my favorite is some sort of distributive come to tation system which one of our customer has built so once you a so one thing needs so they a top four o'clock softly called silvers and so undeveloped kind of execution by flying drones and the date type shit or stock market data in but in general in all stores and only a dollar fine to settle that's a to to do very complex mathematical company stations stations that comes with not having to be quite position it calls an auto are all partial lifeline i seniors a data in a in a together in the single table to be accessible by by the users you've sprinted easy to a to develop a sense income sparkling hunter but it would take a so these a florida dot com here's in a given a data scientist at eight deals include carlos so just so hours until.

kodak kurt six seconds four days
"coober netties" Discussed on Windows Weekly

Windows Weekly

04:14 min | 3 years ago

"coober netties" Discussed on Windows Weekly

"So what's new what's actually going to come in nineteen? Oh, three. It's I would say a bunch of incremental things they said they focused on four core. Pillars at platform the edge computing stuff edge with a lower Casey windows admin center, which is the project Honolulu graphical user, interface stuff an app compatibility in the blog post. They talk about each of these areas and some of the specific features that they have added to serve nineteen three and the one on this list that surprised me the most is we finally have windows container support in coober netties. I thought that had been there for ever. But I guess it was just there in previews. And now it's there in g a form and they've now integrated support with that with nineteen three on edge on the edge compete. Eating stuff. They they're working on making server work better with specific edge scenarios admin center, they continued at all kinds of new features people love that Edmund center, if you haven't if you've never tested that out you should give it a world and they've done a lot of things to improve server core app. Compatibility? They've got lists in this blog posts. Also, you can get things like server core insider builds on containers, Nanno server container builds and windows container builds for insiders. So it's a post worth checking out. And finally, we know a little bit about server nineteen three that's going to be released in late may alongside windows. Ten nineteen o three. So that's my pick some good co night codename here if you're trigonometry. I know this is a super interesting codename one I had not heard until this week. And now I've heard it twice this week code him. Cosign and cosign stands for core s an intelligent edge again edge in this case, the lower case E, very funny. So what is this cosign is windows core? When does we keep talking about windows core OS, and what's going on building out windows Koro s the team that is building. This is known as cosign at Microsoft. And so we've we've heard through the grapevine that we're not going to hear about one does light at build next week windows light is Microsoft's chrome OS competitor. That is built on top of windows core OS, we're not going to hear about it yet. But the fact that we now know the. Name of this team. And we also know how lens to uses C O S as well makes me think at some point fairly soon. Microsoft is hopefully going to start peeling back. The covers on what they're doing with Koro s and to me most interestingly how they're gonna get win thirty two apps to run on this specifically on windows, light word is through containers virtualization. But maybe the cosign team we'll have more to say about that at some point soon. Excellent. I don't know about you. But I feel we've all earned a frothy beverage a frosty. Cold, frothy Hat tro. tro is yes, yes. For paul. It's a rose as slushie. Over. But okay. But my pick for the beverage of the week apples to be a beer and happens to be a beer made in Seattle since we're all going to build next week. There's a Seattle berry cult holy mountain, and wholly whatever are they good. They make a lot of on style beers. They also make IP as and other dark style as well. But they make excellent Cezanne's including my pick, which is called witch finder. Which finder is a beer that's kind of a mixed beer. It's for mental with Brett, which is a wild yeast and that gives it a slightly sour Tang. But also a refreshingly sour Tang. So it's a Cezanne may with Brett it's only like five percent. You can find it on tap not just the holy mountain taproom, which you should visit..

Microsoft coober netties Edmund center Brett Cezanne Seattle Honolulu paul five percent
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

04:10 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"Bonetti, it can be Amazon ACS very soon gonna be supporting dockers, warm or hassle. Nomad any scheduling platform four containers that you wish and then once as contemporary scheduled the computers, basically abstracted away from you equal scale it will kill zero will scale to whatever copies of of of computer, you need automatically. You don't have to do anything in donate to manage the M's jumped by security patches or fetches a already done in the background by by ocean. And basically gives the developers service way you manage their containers in also very very reasonable pricing of computer, they can on lock as they scale the experience of deploying a container on your product. Ocean. It seems quite similar to deploying a container on an AWS far gate or as Azure container. Instance. These are these long-lived container instances, the server service containers, how do the containers that you're spinning up for people on ocean compared to these container instances that the cloud providers offer. So I think we're hitting a spot where we basically offer the same service like like Ozzy container instances, the or. And a just a matter of like how efficient we are in acquiring the underlying computes and our goal is to be way more efficient that any other provider in acquiring the compute underneath of the container. Right now, I find these container instances so interesting and so useful because I've been a longtime user of horr- Oku and her Roku is as far as I know just giving you a container it's been giving people containers for a very long time since before containers were trendy and they're called Dyno's on her Oku and heroic is made a great business out of selling Daigneault's to people because that's all you need often. You know, many people who are hobbyist developers or even at a large company and they wanted to use her Roku. They can spin up their entire application is just some monolithic application that just runs in a long lived container. And and that's great. That's what a lot of people need. So that's what AWS and Azure have been offering with this far gate and container instances products respectively. And I was talking to you about this. When we were at qb con, but it surprises me that there's not more excitement around the idea of these low cost cloud provider container instances where you can just spin up your monolith in standalone container people see more excited about the idea of managing a Cooper netties cluster. When in fact, like why would you wanna be a cluster operator? You could just be a an operator of your application. Abstraction? Why is that why do people want to manage their coober netties when they could just manage their application in terms of these container instances, you know, different. I couldn't agree more. I think you right in terms of also its greatest you how what he Roka did like becoming so trending now. And also if you go like that and look at Google app engine, which is basically the same thing. Equal. That's true. And ready for not yet. So and I think this is like the strength of eight of US, which they are actually meeting the market, whatever the market needs death. So they will not introduce like, you know, some spaceship when you you just need like a regular car. And I think that that's a big thing for Amazon and right now, he's becoming popular because people like did that you know, runway. Like, hey, we knowing what is container. Okay..

dockers Bonetti Roka US coober netties Amazon horr- Oku Ozzy Google Daigneault Oku Dyno
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

03:46 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"Be Amazon E C S, very soon gonna be supporting dockers warm, or or hassle nomad any scheduling platform for containers that you wish and then once as contemporary scheduled the computers, basically abstracted away from you, you will scale it will fail zero will scale to whatever copies of of of computer, you need automatically. You don't have to do anything in donate to manage the M's jumped by security patches or fetches a already done in the background by by ocean. And basically gives the developers service way you manage their containers in also very very reasonable pricing of computer day can on lock as they scale the experience of deploying a container on your product. Ocean. It seems quite similar to deploying a container on an AWS far gate or Azure. Container. Instance. These are these long-lived container instances, the service containers, how do the containers that you're spinning up for people on ocean compare to these container instances that the cloud providers offer. So I think we're hitting a spot where we basically offer the same service like like, Azure, container instances the or fight it. And and just a matter of like how efficient we are in acquiring the underlying compute and our goal is to be way more efficient that any other provider in acquiring the compute underneath of the container right now. I find these container instances, so. Interesting and so useful because I've been a longtime user of Hiroko and Hiroko is as far as I know just giving you a container it's been giving people containers for a very long time since before containers were trendy and they're called Dyno's on her Oku and heroic is made a great business out of selling Daigneault's to people because that's all you need often. You know, many people who are hobbyist developers or even at a large company, and they wanted to use her Oku they can spin up their entire application is just some monolithic application that just runs in a long lived container. And and that's great. That's what a lot of people need. So that's what AWS and an Azure have been offering with this far gate and container instances products respectively. And I was talking to you about this when we were at qb con, but it surprises me that there's not more excitement. Around the idea of these low cost cloud provider container instances where you can just spin up your monolith in standalone container people see more excited about the idea of managing a Cooper netties cluster. When in fact, like why would you wanna be a cluster operator? You could just be a an operator of your application. Abstraction? Why is that why do people want to manage their coober netties when they could just manage their application in terms of these container instances differ. I couldn't agree more. I had you right in terms of also its greatest you how a what he Roka did like becoming so trending now. And also if you go like that and look at Google app engine, which is basically same thing. But just that's true and ready for that yet..

Hiroko dockers Oku coober netties Amazon Google Roka Daigneault Dyno
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

04:42 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"Engineering daily dot com slash Mesa's. Fear. Dan, con-, you're the executive director of the CNC f-. Welcome back to software engineering daily. Thanks, Jeff happy to be here. I had a great time at qb con in China. And we spoke there at some length, and what was interesting about our conversation was some of the reflections on Chinese software companies and how they compare to those of the US give me your perspective on how Chinese software companies compared to those of the US. Sure, I think the biggest thing that China, and particularly the caprice that are focused around software have going for them is that they don't have all the history that we've built up in Europe, and and the US, and so the biggest example of this is that they didn't spend one hundred years investing into fixed line telephones, the jumped directly to mobile phones. And so that's meant that their usage has skyrocketed, but also that ubiquity has enabled things like mobile payments. And such that in many ways are far ahead of the US Europe on the software company side. I think there's a a newer phenomenon where a ton of companies just don't have the same ten or fifteen year investment into virtualization. They kind of skipped right over the previous generation of technology. And so a really in a position to adopt, cloud native, much, faster and are doing exactly that this is a leap frog effect. I think you called it in China. What are some of the ways that the leapfrog effect impacts how these Chinese software companies operate? Yeah. The other term will use this second mover advantage that coming later actually can give them a big heads-up oddly enough. And it really is just an extraordinary thing. I I actually made six trips to China in twenty eighteen and a number of the year before the year before that and comparing it to my. I I visit there in nineteen Ninety-seven. It's just a completely different country. Everything about it has changed in. When you look at these modern technology companies by do which many people see as the Google of China and a ten cent, which is kind of Facebook and Twitter of China, and then Alibaba which which plays a similar role to Amazon both on the retail side. And also with the cloud services that they're just moving incredibly quickly in and so eager to adopt technologies that have come out of the west, but one of the really exciting transfer. Now seeing is that they're also a very eager to contribute technologies back that that really makes it a global open source community in the United States. We know the dynamics of the cloud providers. Most people listening do Amazon is the biggest and then there's Azure, and then there's Google. And then there's everything else. Tell me about the cloud provider market in China. It's definitely different. So you have this a great firewall of China that the government controls a huge amount of the content. That's coming in and out of the country. And that means it's extremely unreliable to try and reach Chinese consumers from websites that are set up internationally. Now, you often can do it. I mean. CCF dead. I o is an example is one that that they often are able to reach a reliably. But one of the things that Cooper, Netease, CF has been able to do for a project is we just set up a content delivery network with Alibaba cloud inside the great firewall. So that copies all the content from coober Nettie Stilo to communities dot COM and that a Chinese site is just a far far loads far far faster inside China, but from a high level Alibaba has a similar role in that they're the dominant cloud company in in China to what Amazon plays in the US. And then the other major companies like a while way and ten sent in by do have cloud offerings. And then there's a number of smaller ones as well. Like Nettie using easy stack and and many others. But Alibaba is definitely the eight hundred pound gorilla. They're possibly the largest Cooper Netease. Clutch. Her in the world runs in China at JD dot com. Have you learned anything unique about coober netties from seeing these large scale Chinese deployments?.

China Alibaba United States Amazon cloud Europe Cooper Netease Google coober Nettie Stilo coober netties executive director Dan Nettie Jeff CCF
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

04:30 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"Auto scaler didn't get released until well after initial launch and the initial guidance around that was well, it's it's not the first to market. It's not the best of breed. I'm not in. Nearly sure why I would use this today that said it has in typical Amazon fashion improved rapidly since the time of launch to the point where now looking at it. It's not at all about option that said, I think the larger ecosystem story here is that coober netties, and or how we orchestrate containers is not going to be an area of focus for too terribly long. I mean, when's the last time you had a deep dive discussion with someone about which lyrics distribution, they should use. It becomes plumbing. It starts slipping below the surface and stops mattering to most people. I think that container orchestration is absolutely going to follow suit. And what's interesting about that is you know, what has gone from the server 'less idea, the Lambda AWS Lambda has gone from a fringe kind of cool idea to looking like that's going to be the modality that people are building their applications in from day. One more and more going for. Although it's hard to to see exactly what the server 'less stack will look like. Because today, you, you know, I guess the the most server 'less stack that you can build would be something where you're using a bunch of managed services like queuing and database some have database system. And then you glued together these big managed services with glue code that runs in AWS Lambda. But that could certainly change you could certainly have further developments of things on top of Lambda. What's your sense of the adoption of server 'less? I think that service. This is one of those interesting technologies that at what it launched. It started to look like an awful lot like a toy where it's the sort of thing that you see that more or less manifest itself in a way that okay, this is neat for this one use case. But never for my use case, it only supports some certain languages in it only winds up working for five minutes, and as capabilities continued to. Band, and it starts to move up the stack. It begins to be something that starts to look a lot more realistic. The I think hidden secret behind service powers at is the event model where an event happens in your environment. And it automatically invokes a reaction to it. That is highly parallel Elizabeth, and it almost completely removes the need for a company to look it infrastructure. It's also priced incredibly competitively to the point where there is virtually no company on the planet that is spending a huge amount of money on Lambda. So it it winds up shifting the attention in a few different ways. The problem is is that just be look at service that term has come to mean a lot of things to a lot of people, and it's not helped by people chiming in uselessly to proclaim service still runs on servers as if it were this revelation about not occurred to anyone until the person chimed in and great terrific. Thank you. Do you anything else to add in the same hackneyed comment that? Everyone makes all the time. That's not helpful. And you're not advancing the discourse in any meaningful way. Although there are fair criticisms of Lambda, it doesn't do everything these days. I have a list. Oh, tell me your list. What are the shortcomings of using lamb to these days? The initial problem, of course is shifting the way you think about software into this new paradigm. If you have a twenty year old monolithic application that has been doing something of business value, for example, running the ATM networks, or this is what makes sure the traffic lights. Don't all turn green at the same time, then shifting that to Lambda is likely not going to be beneficial in any strategic sense. Whereas if you're looking at this from the perspective of we're building something greenfield. And this is the way we think about things maybe that adding significantly more value the biggest danger from my perspective is people who see this. New tool and see it strictly as a set of check box capabilities, but don't shift the way they invision architecture where they still try and shove and entire monolith. Into a Lambda function where they wind up try to store state inside of Lambda, you can build an awful lot of terrible things with terrible anti patterns using something like this technology..

coober netties Amazon Elizabeth five minutes twenty year
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

03:19 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"Give yourself the ability to actually do that for all the other services because if then, you know, you do find yourself using a Google specific service and twelve months later, you're no longer running that application or that application runs just as well consuming GP's in a cheaper way on Prem or. Or in some other cloud rather than having to do one hundred percent of work to completely move yourself. From one cloud to another cloud. You're doing like ten percent. You know? So there's some work you still need to do because you did tie into some particular API's. But for everything else, you know, you're just you're you're just using the open source stuff. So again, we've got customers that we work with where they'll use our coffee. They'll use our spark they'll use our coober netties, they'll use are elastic, and they'll use then one service from a particular cloud that they want to use that's critical for them to use. But they've still given themselves a lot of leverage in a lot of opportunity and just as much simpler time if they ever want to move in the future to be able to do that move if they hadn't done that. If they would've built their applications directly against all of those services in the cloud that would have been tough, you know, anything about move that they wanted to make I would have been a very very long effort. It's unclear that they would ever been able to. Do that effort. And it's unlikely that they would. But we have customers that have moved apps hundred percent from on Prem to the cloud because they've been running on top of the platform. That's powerful. It's really really really powerful. So it's very very possible. Yes, you have to move data data has gravity that that can be a real thing. But but people can do it. They can do it. And I personally think that in the future as multi cloud becomes more and more thing, they'll be price worse. You know, they'll be there's already price worse. But I think you'll see this be a real real thing. When customers are able to actually apply some some leverage and say, hey, listen, I can get a cheaper over here. And so I'm going to do that. And I think that's a good thing. I think that's a good thing for business. So if I want to interact with an Amazon service today, and I want to be like an Amazon specific service Amazon, for example, and I want to have my coober Netease cluster interact with it I've got some application that wants to make make calls out to the Knesset API. But I want I want Mesa fear to be my multi cloud management layer. So I've got certain concious calls. I wanna make in my application. Maybe I've got certain Google big query calls. I wanna make what if I'm using Mesa sphere as my platform. Do I need do. I also need like an instance in Amazon do I also need an instance in Google or do I just make calls out to the API? It was that low. Yeah. You can you can just make calls out directly to the API is there's there's two parts to to this magic. The first part of the mad. Magic is how you get the services them selves. How you get instances of the services because even with you know, the. Nieces or the big tables of the world you still to hit API's to give yourself to provision the service. So that you can then make API calls to it. So that that's kind of the first part and the direction that that that.

Amazon Google coober netties one hundred percent hundred percent twelve months ten percent
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

03:05 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"Take advantage of it and make it part of your company or you wouldn't yourself be be the ones that are doing the innovation to be in the ones that are doing that. And I think businesses know that and that's why they create these research arms, and there's a bunch of really really good tech that can come out of those research arms, and that ends up being a balance. But again, I mean, if I go talk to some of my customers, and I say, hey, I know you want copy and paste, and it's gonna make your life so much better. But like, that's know, kind of par for the course. So we're not going to build that. But we're thinking about this crazy other future blink copy. You know? I I know what my customers would say. They'd say, whoa. Whoa. Whoa. No, no, just just give me the simple thing. Just Jimmy Ryan faced leading a perfect example. This is where you've gone with the company because you start out with something that was. Revolutionary base providing cluster scheduler data center operating system stuff for people, and you found traction there and the market has gone to a place where the focus of enterprises today as far as I can tell is that they want an ability to manage multiple open source frameworks, which does fit into your original vision. But there's also the focus on how do we manage a Jillian coober netties clusters? So it's a focus on Cooper Netease. And it's also a focus on how do we adopt cloud services and multiple cloud services from different providers? And how are we going to manage all the billing, and how are we going to manage all the different accounts? And it's just kind of a disaster right now people have no idea what to do. If you go and talk to different vendors who are selling to them, the vendors can give them some sense of vision. But the divisions are different if you from vendor to vendor so you've enterprises walking around, and they're kind of you know, some of them are a little bit paralyzed. Some of them are. Like gradually easing into their buying decisions. So tell me a perspective there. How do you position yourself in this super crowded market or you have all these different kind of distributed systems modernization vendors trying to compete for the enterprise contracts? All the vendors are offering different visions. How do you position yourself? Yeah. Yeah. I think you're actually capturing it pretty well for us, which is it tons of Cuban is clusters, but not just a bunch of other services distributed systems data services people are trying to run as well thrown into a mix of do. I do this myself or do I do it myself on Prem or do? I do it myself in the cloud or do. I just let the cloud. Do it all for me. It's actually it's it's a pretty fun time from from a business perspective when it comes to that. Because I mean, you'll even hear this from for WS and the leadership eight of AWS, we're still it's still early days for cloud. It's still really is. There's a lot of organizations that are in the cloud, but it's still early days, and there's a huge opportunity. To help people with that transition. The thing we focus on all actually actually give you a vision for the company that you know, that we say all to all new employees. And it's kind of our elevator pitch in many ways. What we wanna do is provide public cloud experience or public cloud services from.

cloud Jillian coober netties Cooper Netease Jimmy Ryan
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

03:10 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"That your customers are going to have available to them, and you can provide a set of plug ins that will allow your Cooper, netties instance, to be able to communicate with those volume. Begins and prior to see aside, those plug ins were built into the core of coober Eddie's. So the challenges were in, you know, figuring out the bugs in those storage systems and the bugs of the coober Netease layer. But there wasn't there aren't inherent major complexities there. If a particular cloud customer chooses to use other storage on top of what is provided by the cloud provider for the most part that is the responsibility of the customer in the storage vendor that they choose on top that provides an additional software defined storage layer to figure that out. How are you going to deploy it how is it going to interface with Cooper Netease, the challenge I think is for folks trying to run coober Netease on Prem in their own environments and trying to figure out how do I run these storage systems? How should they interact with coober netties who is going to provide me, my CSI driver. How do I get it running? If I'm running within a virtual is environment. I'm running VM ware or something. How do I expose storage through? That VM layer into the into the the VM. So that they're available to the containers inside those at least for our specific cloud provider. G E, Google cloud. Guber Netease engine where we recognize that problem, and we're trying to provide a managed solution for on Prem as well where we'll say we're trying to solve a lot of the problems that exist in that space and give folks there a managed solution as well. All right. Well, we're here at qb Khan, and there's a gigantic amount of people here. Seventy five hundred people you work on g k so you're at the epicenter of Cooper, Netease development as much as anybody give me some predictions or tell me what you're excited about in the Cooper Netease community. That's a good question. I am very very excited to be here. Tim Hawkins who was the person that recruited me to the cougar Netease team early on in two thousand fourteen likes to say that when I'm at q Khan, I feel like I'm with eight thousand of my best, friends and. Honestly, I feel that way too. It's it's very nice. It's a great wonderful community in terms of what I'm most excited about. I think there's a couple of things one is just continued adoption and growth of coober Netease in realizing that we're reaching a state where a lot of the innovation is not necessarily happening within coober Nettie is it's happening at a layer above coober, Netease and figuring out how we enable that. And how we bridge the gaps. So that this ecosystem continues to grow. And then the second is basically an extension of that. Which is what are the kinds of features that we're going to expose there that are going to make -application developers lives easier and make deploying and managing applications in clustered environments dead. Simple the way that I like to think about it is that, you know, back in the day before there were operating systems folks who were developing applications had to be in Neethling aware of the specific hardware that they were deploying that particular application on and..

Cooper Netease Netease coober netties coober Nettie Cooper Khan CSI Google Tim Hawkins Neethling
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

03:10 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"That your customers are going to have available to them, and you can provide a set of plug ins that will allow your Cooper, netties instance, to be able to communicate with those volume. Begins and prior to see aside, those plug ins were built into the core of coober Eddie's. So the challenges were in, you know, figuring out the bugs in those storage systems and the bugs of the coober Netease layer. But there wasn't there aren't inherent major complexities there. If a particular cloud customer chooses to use other storage on top of what is provided by the cloud provider for the most part that is the responsibility of the customer in the storage vendor that they choose on top that provides an additional software defined storage layer to figure that out. How are you going to deploy it how is it going to interface with Cooper Netease, the challenge I think is for folks trying to run coober Netease on Prem in their own environments and trying to figure out how do I run these storage systems? How should they interact with coober netties who is going to provide me, my CSI driver. How do I get it running? If I'm running within a virtual is environment. I'm running VM ware or something. How do I expose storage through? That VM layer into the into the the VM. So that they're available to the containers inside those at least for our specific cloud provider. G E, Google cloud. Guber Netease engine where we recognize that problem, and we're trying to provide a managed solution for on Prem as well where we'll say we're trying to solve a lot of the problems that exist in that space and give folks there a managed solution as well. All right. Well, we're here at qb Khan, and there's a gigantic amount of people here. Seventy five hundred people you work on g k so you're at the epicenter of Cooper, Netease development as much as anybody give me some predictions or tell me what you're excited about in the Cooper Netease community. That's a good question. I am very very excited to be here. Tim Hawkins who was the person that recruited me to the cougar Netease team early on in two thousand fourteen likes to say that when I'm at q Khan, I feel like I'm with eight thousand of my best, friends and. Honestly, I feel that way too. It's it's very nice. It's a great wonderful community in terms of what I'm most excited about. I think there's a couple of things one is just continued adoption and growth of coober Netease in realizing that we're reaching a state where a lot of the innovation is not necessarily happening within coober Nettie is it's happening at a layer above coober, Netease and figuring out how we enable that. And how we bridge the gaps. So that this ecosystem continues to grow. And then the second is basically an extension of that. Which is what are the kinds of features that we're going to expose there that are going to make -application developers lives easier and make deploying and managing applications in clustered environments dead. Simple the way that I like to think about it is that, you know, back in the day before there were operating systems folks who were developing applications had to be in Neethling aware of the specific hardware that they were deploying that particular application on and..

Cooper Netease Netease coober netties coober Nettie Cooper Khan CSI Google Tim Hawkins Neethling
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

03:45 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"Nettie says, okay, there's another controller. So for example, like a replica set controller or a demon set. Controller a workload controller stateful set. Controller for example, will say I expect there to be maybe three pods running on this cluster for this particular stateful set. But I see that. There's only two one of them has crashed. I am going to create a new pod to ensure that there's always three running because that's what the user requested, and it will create the new pod object. When the new pod object is created. It is referencing the PVC that already exists in the system and the existing independent of the original pod original container, and the PVC is the pointer to the state that has been persisted outside of the cluster somewhere. So now when the pod get started, the coober netties controllers work to make sure that. That storage that that PVC is pointing to go at gets attached and mounted into that new container inside that new pod. And when that new pod starts to run all of that state is available inside that pod. You touched on the operator pattern earlier. Can you explain the operator pattern and more detail? I'm sure it's definitely not my expertise, but I can touch on it. It's my understanding is that if you have a complicated app. It doesn't necessarily need to be stateful, but stay Phillips tend to be more complicated doing operations like basically day to operations things. Like, I want to upgrade my application. I want to scale out my database, I want to scale in my database, those require a lot of careful orchestration. That is above the coober Netease layer, they are this orchestration is unique to the application. And so what these operators are there. Just like every other component in coober. Nettie is there is a piece of API which is. CRD which allows a user to be able to interact with this operator. And the operator itself is just a controller which is a pod with an application. That's monitoring these API objects and doing something. And so the idea is you deploy an operator and operator starts to run. You create a CRD that defines I want my application to look like this the operator sees that CRD and says, okay, the customer asked or the user asked me to start a reticence instance, and these are all the configurations specifications for this instance, I am going to translate that into what is required for running. That instance, on coober Nettie, so it will automatically provisioned the workload objects and the PVC objects that are required to get that instance, running once that instances running users can then create more CRD's to do more operations against that operator. So for example, if they want to upgrade to a new version of lettuce they can. Modify configuration file to say instead of res- one point. Oh, or whatever I want to move to one point to the operator sees that change, and it orchestrates the update of the reds instance, by, you know, carefully deleting and updating the coober Netease components that make up that instance. So in that way, it's basically a layer above coober Netease to manage complicated. Applications you work at Google on g k you're also on the sig storage. So that's the the special interest group that is focused on open source efforts in conversations around storage. So you have a lot of conversations with both practitioners and the practitioners at Google and practitioners elsewhere around how they're working with storage, the frictions they're encountering the challenges at the cutting edge..

CRD coober Nettie Nettie Google Phillips
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

01:42 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"And if there's a bug you gotta wait and follow the massive Kuban Netease released process, and maybe they don't wanna open source code in some cases, so together with the vendor community in Kuban Indies at that time wasn't super mature. So what we did was. We partnered with other cluster orchestrators cluding Mazo cloud foundry Docker Docker swarm at the time. We all got together and decided we're going to create a standard called the container storage interface and the purpose of this was to have one standard for how a storage system. A storage vendor? Could plug their storage into a cluster. Orchestrator a cluster orchestrator being coober netties Mazo Docker swarm, cloud foundry and that works started about two years ago, and we're very happy to announce. That CSI went one point. Oh this last month and coober Nettie is one point thirteen pushed CSI support two G A. And what this allows now is that vendors can basically develop new extensions news for new storage systems independently of the Kuban Eddie's release the way that they develop them is basically just like any other coober nineties application. It's containerized there is an interface that's defined they implement that interface, and they deploy on coober Nettie is using the Kuban as primitives that exist just do acute cuddle apply. And now support for a new storage system exists within coober Netease and Kuban eighties is able to allow application developers to begin to use new new kinds of storage without having to modify the core kluber, Netease code. So if you're a cluster administrator previously, you were limited to the set of storage plug ins that shipped with coober nineties, you were of course, responsible for setting.

coober Nettie coober netties Mazo Docker Kuban Netease Kuban Kuban Indies Kuban Eddie CSI Netease administrator two years two G
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

01:42 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"And if there's a bug you gotta wait and follow the massive Kuban Netease released process, and maybe they don't wanna open source code in some cases, so together with the vendor community in Kuban Indies at that time wasn't super mature. So what we did was. We partnered with other cluster orchestrators cluding Mazo cloud foundry Docker Docker swarm at the time. We all got together and decided we're going to create a standard called the container storage interface and the purpose of this was to have one standard for how a storage system. A storage vendor? Could plug their storage into a cluster. Orchestrator a cluster orchestrator being coober netties Mazo Docker swarm, cloud foundry and that works started about two years ago, and we're very happy to announce. That CSI went one point. Oh this last month and coober Nettie is one point thirteen pushed CSI support two G A. And what this allows now is that vendors can basically develop new extensions news for new storage systems independently of the Kuban Eddie's release the way that they develop them is basically just like any other coober nineties application. It's containerized there is an interface that's defined they implement that interface, and they deploy on coober Nettie is using the Kuban as primitives that exist just do acute cuddle apply. And now support for a new storage system exists within coober Netease and Kuban eighties is able to allow application developers to begin to use new new kinds of storage without having to modify the core kluber, Netease code. So if you're a cluster administrator previously, you were limited to the set of storage plug ins that shipped with coober nineties, you were of course, responsible for setting.

coober Nettie coober netties Mazo Docker Kuban Netease Kuban Kuban Indies Kuban Eddie CSI Netease administrator two years two G
"coober netties" Discussed on Software Engineering Daily

Software Engineering Daily

04:49 min | 3 years ago

"coober netties" Discussed on Software Engineering Daily

"Okay. So if I'm building my application on AWS today, even if I'm in coober netties, and I'm thinking, oh, I've got portability. Now, I'm not locked in. Now, if I use dynamo DB Denno DB, I believe has proprietary is. There's no open source like I can't just take that code. And and just have it run in an open source product because it's just it's gonna run better on against dynamo. DB and dynamo DB has donate PI's. But if you're using cross plane, and you say, hey, cross playing like I want you say I wanna keep you store on this cloud over here. The the API's will be more open, and they what they just translate into dynamo API's no-no. So so that's a very good point. So so there's there's no silver bullet here. If you're using something like if your application. Has built into it. The, you know, a client library for dynamo DB. Then you must use dynamo DB's medicine service. You can provision it through cross plane, but on the wire the data plane the data path to the actual service is still Dinamo Dinamo the right? So there if you are using proprietary databases and software from given cloud provider than you don't get to achieve a high level, portability. You can still use cross plan to get universal API Anna common way to kind of manage and provision infrastructure across the different cloud providers. But for in this example, you're still using given cloud providers proprietary stacked pirated software. So you're saying if I'm using my sequel, for example, since it's open source, right contracting example, it's either open source or has a a widely adopted wire protocol, so you'll see. See, for example, that people have normalized on using my SQL or post grass as a wire protocol for many Implementation's of databases. So your choices. There are you know, Maya Grell post grass Aurora, the Vitesse, even cockroach db has a post crest y supports the post gusts via particle. So you get they're different ways of supporting in using databases and still not having to change your application when it's consumed different implementations of it. And with the trajectory that we're on more and more open source software is becoming the default. And so we think that we're moving towards world where there will be more managed services based on open software or managed services that support open or widely adopted wire protocols. Okay. So the advantage there is I let's say I'm getting started with my application, and I write I'm using my sequel backing for. My database and then over time. I'm like, oh, actually my application has such a high volume of data that I wanna use Vitesse. You know, it's totally fine. Because you've still got the the right API's. You've still got the same API's or if you would say, oh, I want global consistency and super resiliency, I want cockroach DB you've got the same protocol. And this is in contrast to to what so it on AWS today, if I was just going with AWS, and I was using Aurora are there specific API's that like specific my sequel, non my sequel API's, no, so you can run an application that just uses a Maya scale client library, and it would run against Laura or run against mice askew or Iran against Vitesse unchanged. Right. And so that's partly the property that we like when are we thought about when were you know, designing cross plane that the decision to use a roar? Oh or the decision to use Vitass or the decision to use mice quell should be a late bound Runtime level decision. Not have Time Engineering decision. And so if you build your application, and this is this is not new in fact, if you look at we were talking about Cabrini zone if you look at coober Netease when I build my application put in containers. I don't think about where it's going to run that. That's it's it's a similar thing. Or if it's using a storage volume. I don't really think about whether it's gonna come from e b s or from Google persistent disk, we think that there's a similar property that could be done for a wider range of resources and services that people are running in cloud today, and that trend is increasing where seeing more and more cloud providers offer managed services of open source software,.

Vitesse AWS coober netties Google Time Engineering Vitass Iran Laura