Cloud 2030: Shaping The Future Of Cloud Computing | Rob Hirschfeld

hi this is your host abner bharti and welcome to tfi insights is sponsored by reckon it's a show where we deep dive into cloud native technologies and today we have with us once again ceo and co-founder of rec and rob hirschfeld and today we are going to get some updates on cloud 2030 just for a quick reminder to our audience what exactly is 23d cloud cloud 2030 um the website is the2030.cloud uh is a set of weekly discussions where we talk about the future it's a very unique session it's it's very open it's very hallway track type formatted and we look at how are we building technologies and infrastructures for the future what are the impacts concerns what's going to change it how is the market going to evolve and with an eye towards can we influence what's being built and do we like what's being built so let's talk about some of the really interesting topics that you folks are discussing at uh uh cloud 2030 one is uh sas economics which is where we talk about capex versus opex up so so let's talk about get into this because ownership cost of ownership does become a big topic when you look at one or two years is fine but when you look at 10 15 years that becomes a serious topic the thing that really jumped out to me in the discussions was that you know everybody everybody has this idea that we prefer capex over sorry opex over capex meaning incremental investments pay-as-you-go type models and and that's pretty hardwired in people's expectation at this point but we went deeper because we were looking 10 years out and the idea here is that that model of acquiring technology influences how we plan to build innovation and the the way we're actually buying incremental technology and not needing the startup capital not needing to build infrastructures to innovate changes the way we expect innovation to be built and so the thing that was fascinating in 2030.cloud is that we really impacted how the systems are built in the future like we're relying on the fact that we can get incremental ads to technology and build things incrementally that we don't need the startup costs and the startup capital and so we were we were excited and also concerned which is a theme for for the discussions that we've really accelerated the pace of innovation but made it in a way that we have all of these dependencies on cat-back sorry op-ex infrastructures and that can be a challenge based on the way we've traditionally innovated where you invest and that investment pays off over time the investment isn't necessary in a lot of cases anymore we also talk a lot about complexity especially if you look at the cncf land escape there are so many logos you need a really 32k monitor you cannot do any more in 4k it's good to have so many choices but it also leads to a lot of kubernetes itself needs a lot of knobs to be turned but when you look at all the projects that are there so uh we're talking about you know java's complexity paradox so first of all tell me what is that before that you if you can a little bit just give us you know how complicated cloud native landscape is today this concept uh comes up over and over again and we we talk about how complex things are and the the rate at which complexity is increasing and you can see that in the cloud native landscape you can see it in the way we're building a lot of these pieces the thing that's really interesting is in these discussions people don't seem to get as concerned about complexity and interdependency in systems as you would expect and we we ex we now predict that systems will get increasingly complex and when we step back from that we start scratching our head why is complexity not adding cost normally when things get more and more complex they get more fragile they get more expensive they become harder to maintain but what we've been seeing and expect to continue is this idea that the complexity that we're inheriting is not seen by the people making those decisions so jevin's paradox is a is a paradox where as something becomes cheaper and less expensive people consume more of it think of gasoline with more fuel-efficient cars people drive more because the unit costs the unit economics have actually made it less expensive you get more work we are seeing something similar going on with complexity where the cost of adding dramatic complexity into a system is not being felt by the people building the system think of all of our ai and ml infrastructures where you can add machine learning into an application stack without having to do any of the work to build the machine learning models the algorithms the compute infrastructures you just get it and can add it into your applications that type of additional complexity is very real but it's not being felt by the consumers and the builders so what we're doing is we're making systems increasingly complex at a faster and faster rate and we're not seeing the cost of that complexity that's a jevin's paradox in complexity and so the question that we've been asking is is this a sustainable thing or are we just hiding technical debt in places that people don't know and so when we look towards the future uh the the overall complexity of the systems we're building is expanding dramatically and faster and faster and what we don't quite know is if it's a sustainable pace or not and so there are definitely failure scenarios that we talk about where security or regulation or just the complexity itself with with supply chain fragility could actually send the whole system crumbling yeah one thing i want to ask you this could be off topic is that as in today's world especially because of this pandemic it is quite clear that modern businesses they have to be software companies as well you cannot be business without any software stack you know serving your customers so it doesn't matter what you do and then uh cloud is a critical piece of you know software alone you know you're not running anything on your local computer you're in your cloud now when we look at this cloud-native complexity if you do want to move faster if you do want to build your services and if you look at this complex complexity it's overwhelming it's inter it's it's very very intimidating that is why we see a lot of you know popular day low code or no code where you don't need to know much about what is running under in the you can quickly move faster but the challenge is that if you yourself do not know what services are or what technology is powering your own business you are in a very dangerous situation you should know and right now you may be a small company you're just moving for but as your company grows if you become a target for acquisition and if you're touching a lot of open source code and if you do not even know what is running in your own stack you lead it leads to a lot of security issue it leads to a lot of compliance issue you know where you may be violating a lot of open source licenses sas does make it easier can you can you talk about the balance here because when i look at it i would like yes it enables people to move quickly but you deal with the clients a lot so do you see a trend where companies they get started with no code or low code or you know easy but then as they grow they try to gain insights into the software stack or they are like oh my god what are we doing i'd be happy to this is actually connecting the other two topics right we have this increase in opex generated innovation meaning i can rent somebody selling me a solution and fixing something with the sas and hide all that complexity and yet at the same time as we grow the our dependence on those things the need to understand what they are understand our supply chain protect ourselves from risk those things start to come to the fore and we do see companies that are starting to evaluate their end-to-end systems and then pull back the the way i see this is that things have been evolving and moving really quickly as technologies we don't entirely understand how everything is getting built or what's required or what's necessary and some of those things are going to shake out that we're going to pull back and look at the overall stacks that people are building and then try to streamline them simplify them today we've made a very balkanized market where somebody says oh i have kubernetes but i also need these six products to make kubernetes work correctly six is probably a low number for most people and those things you know over time are going to consolidate into single products and then simplify it and once you've done that then you could actually potentially buy that product add it in and run it yourself and so i expect as this jeven's complexity paradox hits there's a an outcome where people look at that and say this is not a sustainable pace i need to pull back that's traditionally the pendulum swing back to simplicity or it could be that this is actually a sustainable model and and things will go forward um and in that and that uh you know it's that's the harder model to predict because we've never seen systems as complex as we've built um one thing that's been really interesting we've talked about serverless technology quite a bit it's almost surprising to me we get into what i think is a economics conversation or an access conversation or an open source conversation and it gets pulled into a serverless conversation or event driven computing and the reason that keeps happening in these discussions is because that is one of the ways people expect to cope with complexity if i've built my data center my infrastructure and my software stack to consume a lot of these small services that hide complexity i need to glue them together and serverless is what people see as the way to glue or cope with this increasing complexity so this is where it's possible that all this added complexity will be managed for us by having a serverless event-driven system at the heart of all our computing and we'll continue to see that trend rise that's a pretty predictable trend systems are complex serverless helps manage the complexity it certainly doesn't reduce it one thing with serverless is that it's okay for uh relatively smaller companies but if you look at a lot of companies they do a lot of things on prem so even if you do talk about serverless there is a server that they have to manage and maintain so are we talking about only those companies who are going full on sas they don't have anything on prem but if you look at the company who do do a lot of things with opera i think reckon also helps in that space as well so how does serverless help those companies who are managing server either way yeah serverless sadly is one of the worst names for a computing model that the industry could have come up with it's really just event processing so you know there is no doubt and racket works very hard to make servers easier and easier to manage one of the things that we see in cloud 2030 is there is an explosion of innovation going on the hardware side edge computing arm risk new hardware models smartnics right there's a lot of opportunity to improve the physical layers of the it infrastructure we depend on also and that's completely separated it'll end up being hidden as a platform for consumers the end users and application builders but because all these things are being decomposed as services those services have to be connected together and that's what the event systems will do what we would call serverless computing so they're not at all incompatible as a matter of fact that event driven system will be the core glue even more than kubernetes has been or virtualization has been um that you know it's a necessary piece to connect all of these uh disparate software stacks that we're building together so uh when you're talking about serverless or when we are looking at serverless we are now looking at you know the serverless that marketing teams like to call about it it's more or less like how you do things even on your prem as well that is what and that's what you know even even when you say vms and virtual machines you know you're you you are creating those vms on your bare metal you know it's not everything so so it's more or less like the process the approach you're adopting to even even approach right now let's uh let's talk about uh the another topic is uh hyperscale uh the kind of narrative is changing around hyperscale what kind of discussions are going on in that space so one of the things that we see is a general concern about hyperscalers so you know without going into too much detail in these conversations the thing that's interesting is we all expect hyperscalers to continue to be dominant and the key players in the industry and it doesn't take any real um discovery to realize that most people are concerned about the scope of power and control of these companies and then that leads us very quickly into realizing that geopolitical boundaries and supply chains will ultimately influence how these how these companies behave uh and so we do have expectations that as these companies become ever more powerful that other forces will step in and disrupt the their dominance either by creating geopolitical challenges monopolistic breakups things like that and so we do see that the trend line for hyperscalers is going to be very impossible to disrupt without having some type of external influence disasters governments some type of major security or supply chain problem one thing that we cannot not talk about which is ironically kind of elephant in the room right it's actually not elephant which is edge i do remember some discussions around the arm when armed developers they were trying to get a lot of their code into lens kernel and they were looking for energy efficiency and if you look at intel amd folks they're like we don't care we have big massive a data centers but our people were worried about that and if you look at today's world all those codes that was contributed or taken by the arms they are now saving millions of dollars in heat or efficiency in data centers today so those changes do make a lot of impact even you at that point you see hey we don't need it so i look at edge also that edge is also driving a lot of things that you may not consider in you know other use cases so can you talk about the impact of edge whether it's about you know performance utilization talk about how edge is impacting not only the use cases but the stack itself so edge is really this place where we see computing coming into daily life in in even more and more dramatic ways and one of the things that was fascinating uh is that when you think about how we are starting to push technology deeper into our lives and needing to conserve energy and preserve uh power you know supply chains and and balance all of the resource constraints that we we have um and connect to everybody else the need for more and more computing at the edge in our environment local to us will only increase especially with cross-connected systems and coordination and and that's that's a a critical part of life arm is going to be a part of that it's going to be a very heterogeneous diverse system not as not like big cloud providers where everything is the same or very homogeneous or centrally managed and so one of the things that we do see is that trend line will continue to accelerate today it's a bit of a mess and we expect that we will continue to see a lot of evolution in this space but there is no doubt at all that our dependence on computing as a core infrastructure is going to increase dramatically and the 2030 crowd isn't particularly enthusiastic about what we see at the moment for this but we do see needing to have standardization better management being able to have systems that we can really rely on our core components and that's a topic that we ex we will be continuing to talk about in the future because it's a place where we feel like there's some influence where you think figuring out how to connect all these pieces together will make a big difference and i'd love to have people come into the con the conversation and and help us figure this out right we actually can can impact all of these technologies and think about ways in which our daily lives are influenced and can make changes to the future rob thank you so much for sharing these very interesting and very important discussions that are going on within uh 2030 cloud community and and these are also the decisions which are going to influence and drive uh the space the ecosystem itself this is going to be a interesting year last year was a bit where companies are rushing towards cloud native technologies they were moving quickly and now things are settling down technologies like even kubernetes and things that they're mature i mean they are mature but they have moved into the uh deployment phase so this year will be very very interesting how things further consolidate so thanks for this discussion and i look forward to talk to you again thanks swap i appreciate the time and if you're listening please come in join us the 2030. cloud we're having these discussions weekly um they open up to ideas that really just blow me away where we really think about the impacts of the technologies and how they're getting shaped in the future so i'm looking forward to you know expanding these con these conversations and bring in even more people [Applause] you

Channel: TFiR

Views: 1,164

Keywords: #OpenSource

Id: uvELO1qa-7I

Length: 18min 14sec (1094 seconds)

Published: Thu Feb 25 2021