Managing Linux workloads in Azure with latest technical advancement with Red Hat SUSE and Canonical

Show video

[Music] hey everyone donovan brown here welcome to open azure day microsoft azure is the trusted open and hybrid cloud where we offer flexible options for our customers to run any application for their linux and open source workloads our approach is to be open and collaborative and to work with all the partners in the linux ecosystem we contribute upstream to the ecosystem to enable our customers with their cloud adoption and open source implementation today we will be discussing some of the most common questions we get from our customers as they migrate to the cloud we will hear from our linux distribution partners and see how they address these challenges and provide our mutual customers with a smooth onboarding and operating experience on azure let's start with our friends from red hat we have stephanie chilris senior vice president and general manager of the red hat linux enterprise linux business unit joining us today hi stephanie hi donovan it's great to be here thanks for the opportunity so much we know that red hat has a long history with microsoft and we have jointly developed many solutions running on azure from rail to azure red hat open shift could you share the technical trends you have observed when customers are starting their migration and operating in the cloud sure absolutely i think let's start with all the things that are changing really there's so many more options today and they're constantly growing i think the word hybrid which used to mean just on-prem or off-prem now that term hybrid can really apply to multiple levels in the stack the choice of where to run of course on-prem and off-prem still underpin that but now you have options like manage services you have the edge coming in so those options are growing you also have different options of how you run do you want to run bare metal virtual machines do you want to run on a containerized or microservices or even cloud native and across all of that as well now you really need to decide what to run so a devops environment where you can predictably create new applications be able to turn that on in production do it quickly that's never been more important and with all of these changes and options of course equally important is the things that don't change you still have to be able to do this reliably with resiliency and for sure securely and at scale so pulling together things like your management and automation practice that is critically important as well all of this comes together to really underpin what we view as our our open hybrid cloud strategy at red hat it's all about flexibility and choice flexibility and choice with what you run today but also with whatever you want to run tomorrow it builds upon our whole portfolio and then pulls in the ecosystem just as you said the linux ecosystem is critically important it's what brings the flexibility and choice so those partnerships they are essential like our partnership that we have with microsoft of course awesome there's a lot of different strategies and roadmaps that are that our partners are coming up with can you share some of yours that are helping drive our customers to a successful implementation on azure so we have certainly been partners for a long time for over a decade now and i think the strength of that partnership is because we're able to work both jointly engineering also in the upstream but also together as companies and we've worked through really having collaborative support models that allows us to provide a really integrated and interoperable experience for our customers and they are at the end our joint customers and we can deliver to them with trust that will both be focused on their needs and as we have looked at the engineering collaboration we focused on a few things um that are becoming become core of what we want to offer together and that is security manageability performance and that allows us to deliver things that are differentiated both on-prem but certainly in the azure cloud now we're consistently collaborating with microsoft to deliver a more secure and powerful enterprise linux experience in azure this includes support for microsoft features like azure disk encryption and azure role-based access control allowing them to work in tandem with all the advanced features that we put into rel security and we expand that through bringing things in like the library of industry-specific security profiles so that feeds into opens cap or secure content automation protocols and through via enhancements writing key technologies that we feel are core to a linux experience in rel and that's things like security enhanced linux or commonly termed se linux we've taken this further to workloads like sql server on red hat enterprise linux microsoft was the first cloud provider to work with red hat in order to deliver pre-tuned ready to run images in the azure marketplace that incorporated rel high availability add-ons for sql server always-on availability in fact we have similar capabilities are provided in azure for red hat enterprise linux for sap hana so that's really taking that core foundation and building it up into a more workload level experience we've all jointly developed azure red hat open shift as you had mentioned earlier and we deliver that both as a managed service powered by the leading kubernetes enterprise uh expert platform for the enterprise and that's red hat openshift you can choose to deploy that as the managed service or you can install and manage it yourself on microsoft azure we've now expanded that in fact interesting in preview now we've announced the jboss enterprise application platform as an azure app service so you can now deploy jboss by yourself in ad in azure or you can operate it as a managed service you can do that both on rel or on openshift and finally i mentioned management management is key particularly when you start to bridge this whole hybrid platform and red hat management on microsoft azure that can help you unify everything from your physical your virtual and your cloud environments azure arc has enabled support for rel and that allows you to maintain consistent policies and now azure arc has enabled kubernetes and data services for red hat openshift that's currently available in preview so you can see it's kind of the full gambit of what you need to run a hybrid cloud we're looking at all the aspects of what we can deliver on our portfolio and combining that with the azure experience this is awesome and i every time i hear sql server running on linux i just smile so when you said that i kind of how do i because i can't tell you how many times i spun up containers on my machines just so i could get to a quick sql server instance to do some development and it's made my life so much easier so it's just it always makes me smile it's just a different microsoft than it used to be so what advice do you have for customers who are starting their migration into azure do you have any examples of customers who have already done that so what i have seen is that customers are getting real value and leveraging microsoft azure's ability to deliver infrastructure with real agility along with the flexible stable and secure red hat portfolio and we pull this together one great customer example is lufthansa technique it's a great example because they wanted to lever leverage the infrastructure of azure to get that real agility to be able to um leverage microsoft's expertise in deploying infrastructure exactly when they need it but they also wanted a fully open source stack and they were extremely focused on making sure that they had an agile and devops approach and so they chose the red hat portfolio they used red hat openshift as the foundation what it allowed them to do was deploy new applications have them integrate with other applications and they could also lift and shift as necessary when they wanted to pull workloads from the data center up into azure and as they described it they can now go directly from idea to deploying an application without waiting for infrastructure and they stated that there's no limit for them with their hybrid strategy which is just so cool because they can scale with the infrastructure and the platform that we're jointly providing to them yeah it's amazing because when you're talking about the thunza and sap these are the same customers i go and visit right and the stories that we're able to tell as a partnership it just helps us help our customers i mean satya always has obsessed over your customers and i've taken that to heart it's not what's good for donovan it's not what's good for microsoft it's what's good for our customers and it's great to see that you're helping us with those exact same customers and i think we learned so a lot so much from them right every customer that we engage with we learn from them what they want to do next where they're going deutsche bank is another perfect example they were building a new pass platform platform as a service and they called it fabric it ran red hat enterprise linux in some data centers but it also ran it in their public cloud deployment in in microsoft azure alongside all of that they ran the openshift container platform they had allowed them to do container and microservices and both were managed and maintained by redhead ansible tower it gave them a framework to really automate and standardize their i.t at an enterprise scale this allowed them to shorten their development cycle what used to take six to nine months now took two to three weeks that's amazing and they also optimized their cost of their data center because they were able to use cloud capacity microservices containers and cloud bursting in parallel to their data center so that's like true flexibility we were able to combine our core strengths at red hat with your core strengths at microsoft azure and really give them a sort of a lot of flexible options um so those are the kind of stories that really make me light up right this is what ecosystem and partnership really really brings to customers i couldn't agree more thank you so much for your time today stephanie it has been a lot of fun talking to you and and just talking about our joint customers that we work with and i really do look forward to working with more customers together with red hat yeah there's so much more to do nice to talk with you donovan thanks for being here running business in mission critical applications like sap and high performance computing is definitely top of mind for many customers to run these workloads requires deep technical capabilities and fine tuning of the operating system and cloud infrastructure providing a secure scalable cloud for these large applications is key in everything that we do we're honored to have another partner of ours brent schroeder cto of susa to share with us their point of view hi brent hello great to be here this afternoon thank you thank you it's really cool seeing the susa chameleon running some of the largest sap and high performance computing offerings today on azure can you share how you're helping customers optimize their infrastructure when they migrate large mission critical apps to azure absolutely uh you know if we're going to break it into three areas how we like to think about it in the whole notion of simplify modernize and accelerate and there's practical considerations in each aspect of that you know starting off from the simplification standpoint somebody shouldn't think necessarily particularly these large business critical applications as a simple lift and shift and that might work for some small side applications and so forth but when you're going to business critical you there's really several things that need to be taken into account uh such as you know availability zones and how do you set up ha how are the resource management done in the infrastructure and and so those aspects we we think about a lot you know and susan microsoft have worked you know significantly together uh for optimization of that that simplification automation of the configuration and deployment of the infrastructure the configuration and automation h.a you know we think those are absolutely critical you know the most uh error-prone and downtime uh aspects of many systems come around manual interactions and human error and so the more we can automate together then the better it'll be in that migration and evolving to a cloud model versus the traditional on-prem model so that's one of the key aspects and then as as we modernize you know and move those applications then we really want to do more integration together with with how susa and linux work together and tools that and technologies that microsoft has developed for example the azure arc and how azure arc works with susa linux and in providing that holistic view of the linux estate and and wherever you're running a mission critical application you know tying together very often edge and cloud with a mission critical application capturing that data from the edge bringing it up to the cloud and being able to see and manage the entire estate uh through uh azure arc uh is you know a great example of some collaboration that we've done you know that really helps in a mod more modern uh experience that people are delivering and realizing uh and then finally you know as we think about accelerating it's it's how do we help customers deliver faster they've got it up there it's stable it's highly automated and now it's how do we scale with agility and helping customers to adopt cloud native computing containers and the things that we're doing together in containers and also looking forward in how we'll be able to manage uh large container estates in conjunction with virtual machine estates will you know help a customer to deploy applications faster and be able to manage them as they scale up because in a container world you know it becomes very impractical to manage manually the hundreds thousands or even tens of thousands of containers in their inner relationship so as we work together in how customers manage those and continue to build those applications deploy them and scale the applications dynamically uh you know that's some of the the joint work that we'll be doing you know as we close uh some of our transactions in the near future that's fantastic and it's just interesting because as microsoft we run like 50 of our workloads are linux and in the collaboration that we have with our partners has just been incredible in helping us expand that that landscape earlier i know that dr t mentioned that susa is trying to be that bridge between the mission critical applications and azure can you share how you're going to achieve that certainly uh it's built upon what i've already started to to talk about we've got this notion of the the bridge being simplify modernize and accelerate and so that it really starts with step one of automation of the infrastructure uh in in the the deployments and the ongoing day two operations removing as much uh manual interaction as we possibly can and that then brings us to step two or phase two in the bridge uh with modernization and with modernization we want to you know help customers not just adopt uh cloud as a modern aspect but then really starting to move into being able to pick up containers deploy containers in context of their business critical uh applications uh simplify that that adoption as they move to the cloud and start to scale that and then the final phase in the bridge is the acceleration uh and accelerating uh through the use of of you know the more modern tools then also bringing to to the table uh some of our our acceleration through insights you know as you're able to modern or monitor uh the infrastructure better gather more data as to what is going on and be able to respond much quicker to changes in the environment to workloads on the system you know that'll just help uh the customer kind of finish off that transition and bridge to a cloud native world and i like the fact that you're not just sitting on lift and shift because a lot of our customers they really want to get into is but i i tell them and i warn like don't get comfortable there you really start to modernize your applications and start taking full advantage of the cloud and i love i asked it's it's great because it's easy it's familiar you're comfortable with it because you already know what a vm is but it's really not taking full advantage of the cloud and i like how you're helping them take that next step into their journey and make sure that they don't just get to the cloud but they get there in a way that they can start to modernize their applications and use telemetry you know where to make that good investment yeah you know one example if i may elaborate a little bit more on on why um because you bring up i ass and and cloud native and and why we want to simplify that journey not just to the cloud but in in a say a legacy model but taking advantage of everything the cloud and also cloud native computing with containers can can bring to the table i'd i'd associate it with the pandemic today um you know then we all high-tech and and technology always move very rapidly but you know we went through a phase this year that you know really put that in hyperdrive of change happening in in ways that nobody could have anticipated or planned for and you know the i think the companies that are coming out of this the most successfully uh or maybe the least the most unscathed um are those that were kind of already on that journey and and had the muscle memory starting to to occur where they used agile processes microservices containerization so they didn't have to go through and modify you know massive legacy applications to say meet the customer where they need to be met you know if you if you think about you know contactless interaction uh you know curbside pickup at not just restaurants but you know electronic stores home improvement stores every kind of interaction that you had with a customer had to be adapted and if you would have had to make all of those adaptations uh on a monolithic legacy application uh you know i think those are the companies that we see you know that were closed for extended periods of time or may even be now closed permanently yeah because they just weren't able correct they weren't able to adapt and and like you said those were already doing it they just kept doing what they were already doing because they were thinking they're almost future proof even pandemic right because you could they could respond to things that were unanticipated correct correct uh which the cow allows us to do in a lot of special ways one of the other things that i know a lot of people are starting to do and i believe that you're investing a lot in ai and machine learning to bring that to the cloud too to help people with with maintenance of their infrastructure can you talk a little bit more about how you're using ai and ml yeah some of the things we want to do to get some of the places we're starting with uh you know to help our customers you know as people move to ai there's this whole notion of ai ops you know how to how do you turn the operations on and have an operations infrastructure uh for your your artificial intelligence applications and infrastructure for your machine learning and what we really want to do is is deliver an ai platform that has none of the toolkits the frameworks the workflows essentially built into it so that you know from a customer standpoint they don't have to necessarily go searching for which ones fit together you know which tool kit fits with which workflow and then how do i tie my my data scientists into the workflow so we want to help automate that entire estate uh through an ai platform in an ai orchestration environment that's targeted at you know the data scientists and really bringing the ai ops uh members and the data scientists together in in a seamless fashion yeah we're i'm excited about that because we're doing a lot of that too we do ml ops here and ai ops i i lead the devops team at microsoft so you and i have an opportunity to work more together on that ways we can bring that together brent i want to thank you so much for hanging out with us today and sharing with us what susa is doing with uh with azure uh it's been a pleasure talking to you today it's great i've thoroughly enjoyed it thank you and boom2 is known for its openness and convenience in the linux world the rich community of tools operating guides and scripts are some reasons why people love using it for app dev workloads and hosting their web and mobile servers it's the version of linux that i run on one of my laptops we have nicholas demutakis here vp of the field engineering of canonical joining us today as our final speaker welcome nicholas thank you for having me donovan could you share with us how ubuntu is helping customers enjoy the benefits of linux plus azure sure absolutely um if you think of what we like to call the enterprise compute continuum essentially this is compute across hybrid cloud multi-cloud bare metal micro clouds on the edge devices ubuntu has one distribution across everything bare metal vms container base images iot devices like raspberry pi that essentially allows developers to develop and then deploy across all substrates of course including azure in the modern world where workloads need to be sitting in all sorts of environments and serve the end user in all sorts of different ways this is extremely important and the fact that we have one single distribution that helps us uh do this without having a sort of like a paid version and a free version but just a single ubuntu version is the reason why today 1.2 million

machines get started for the first time every single day uh ubuntu is by far uh the the most preferred operating system especially in cloud today ubuntu of course is cost effective right with a single distribution um you can just start any machine or or you can just create build your container there's no license cost essentially right so the fact that you can uh run your cicd start 100 200 machines turn them off or run workloads on ubuntu um without any any effective license fees that has been one of the big big reasons why both startups and enterprises have adopted ubuntu in such a significant way it is optimized for azure we work and have been working with microsoft for years to make sure that ubuntu spins up and runs in the most optimized way in azure both on the is side but also on aks for containers speaking about aks you might not be aware that ubuntu is the operating system for aks workers in azure and so obviously if if people develop uh containers under ubuntu that would seamlessly run on azure's aka services we've integrated uh our own distribution of kubernetes we have two distributions one called canonical kubernetes uh sorry um char kubernetes and another one called microcase for for edge and individual devices with azure arc and so the latest version of arc which essentially manages fleets of kubernetes clusters across all sorts of different substrates works seamlessly with canonical's kubernetes distributions [Music] in azure people can find ubuntu pro which is for enterprise use that offers security and compliance and we also offer of course support and managed services for open source applications across a vast majority of workloads that would run perfectly on azure very cool also i used to be on the azure devops team and i know that our hosted build agents that run linux are also in ubuntu uh agents as well so they're running the exact same version so i i see it all over the place not only inside of aks but inside azure devops on laptops it's just it's been really really cool operating system for me to use um can you share a little bit about the design principles behind ubuntu in the cloud and how it helps our customers build apps with these it seems like you've touched on some of that uh the fact that it's pretty much everywhere but are there any other things that you might be able to add there yeah absolutely um i i brought the subject uh a few seconds ago security and compliance is for us number one priority for production workloads we brought out ubuntu pro which is a version of ubuntu that with a small license fee as a as a tiny percentage of your hourly cost for running vms on azure which enables security patching with slas across a vast majority of apps that you can essentially download from ubuntu's archives and also a large number of compliance primitives uh you can turn on uh with the switch fips compliance you can run cis hardening you can uh run a disaster compliance if you're if you're selling to the us military for example and so on and so forth an extremely important point for us is stability uh stability of of ubuntu running uh in the cloud and of course in all the other substrates that we were mentioning before um and and that has been one of the big reasons why ubuntu is so preferred as a linux distribution nowadays um specifically for azure but also for other other substrates we customize the kernel we have a deep integration with the azure kernel and we ensure that uh the performance uh is exactly the focus of those customizations uh together of course with with all the security elements you you mentioned this before but the majority of innovation uh from from startups uh to any sort of uh large enterprise nowadays is mostly happening on ubuntu the reason uh for that apart from what we mentioned already uh is that there is a vast majority of tools open source tools uh that the community brings out constantly focusing on all sorts of things but nowadays specifically on data science machine learning ai modules and of course this is extremely attractive to developers and eventually to enterprises that of course want to innovate fast the last point i think from a design principle that i wanted to mention is operators operators seem to be one of the latest trends right now especially when you talk about kubernetes aks containers one of the things that developers and devops people have realized is that when you build a complex application which is composed by multiple microservices deploying it putting it together is not very difficult but operating it long term which means scaling out that application creating different relations between different micro services upgrading the application and running all sorts of other operational elements software specifically operational elements it's a very very big challenge to do it with traditional configuration management tools and so as of late the concept of operators has become a trend and for a very good reason and that reason is that operators are essentially ops code that is surrounding the microservice which allow it to be deployed to be related with another microservice to be scaled out to be altered and to essentially be operated long term and so ubuntu offers the largest collection of operators that exists out there in open source because we've been doing operators with charms since 2014 and so we've built all this experience and are now making it available for containers as well as machine operators great so you mentioned developers in your last statement are there any best practices that you would recommend to customers that are thinking about building their applications specifically for ubuntu running on azure sure um some very very high level points uh make sure you build your applications uh as cloud native uh 12-factor uh you have to think in a in a completely different way uh make sure that you use open source uh as much as possible for key components uh for your messaging for example kafka for databases cassandra and so on uh there is a lot of work that goes in these components they're very transferable across different platforms so that could be very helpful you don't have to use containers for absolutely everything it's important to understand that containers are valuable for specific functions but not necessarily for absolutely everything might be easier to just lift and shift certain workloads as i mentioned before use operators try to avoid scripting and of course use apis with your csd both for vms and containers no i love that i love the way that you pointed out the fact that containers are great but they're not the solution to every problem and i have a lot of customers who over pivot there because everyone else is doing it and it doesn't solve a single problem for them but it complicates their scenario a great deal and you can use linux inside of a container but you don't have to there's other ways to utilize that platform for your success so i really appreciate you you pointing that out and i just also want to thank you so much for for hanging out with us today nicholas again i'm a fan of umbuntu i use it myself and it's just really nice to be able to to talk to one of our partners like you so thanks so much it's a pleasure you

2021-02-10

Show video