scala kunal interview 1 26 16
Recording: scala kunal interview 1 26 16
so um hello everybody I'm Alexa crab off the organizer of SF Scala and hilgard location a trap media for our first meetup of the year themed around acha and rapid is very interesting company this our first time here and we have here with us kenapa sim elasticity of offer media and canal how I students color why do you host a scala in the offices and you know what you guys do how scholar helps you do what you do so it's kaveh is a very interesting for us because we are in that you know typical growth stage of you know office startup or piece of technology that a lot of the companies in the Silicon Valley are when they move from you know initial exposure to the customers and proving out the product you're actually getting some rapid development and scale and we build the first version of our platform on mostly on rails as a backhand and you know some call it node and a bench a bunch of other things it's work great for us so far but now as you're expanding to you know amor microservices driven back hand so we can scale individual pieces of our services and our technology our product you know scala becomes more interesting because frameworks like acha frameworks like no spark they're written on top of scala help us get really high throughput streams and you know high throughput event-driven call it messages to our back-end system and that's what we're gearing for you're gearing for like lightweight services that scale over time mm-hmm so the result of kind of different notions around using acha in the enterprise interesting imagined rails because on twitter is this transition first reserves are as epidemic at the fail whale I mean to scale a little dewiness color so so does it solve your scaling problems and on the other hand like is it easy to manage the skull infrastructure is it easy to kind of you know stab the teams what it shows is the city or dressing them so of course challenges there are quite a few challenges you know staffing is never easy as you no one of the motivations of hosting the meetup is to you know get to know some people and for them to get to know rap but more importantly you know we are in the middle of the transition we've just started the transition so we are not using Scotland production today we are in enabling some of our backend office services first and then you know as time goes on our plan is to release you know a few initial services by mid-year which are backed by Scala frameworks I'll call it and based on our experience in building the team and you know how quickly we can expand the ecosystem of our platform will decide how broadly we can use it over time all right solution spark are you guys in this space how do you see kind of you know what services market services connecting to to spark what your thoughts about date architectures like how how far you'd all this kind of planning process so right now our data architecture is very simple you know we have maybe two or three data sources but as we are moving towards more you know highly call it sophisticated analytics back in because a lot of work we do you know the real value is in the analytics how people use wraps and down the road we also see a lot of use cases where we want to do some predictive deployment of content based on the user like the best content of the user wants at the right time of the day and for things like that you know you need massive distributed execution and that's where a spark really comes in so once again you know these are early days but we we are moving towards a polygon data model we all know you know everyone who's done a lot of work with data knows one-size-fits-all doesn't work you know the big vendors would like you to buy one big box yep that does everything like hana or or the exalogic or exit platform from oracle but you know in in our world small autonomous services that scale is the way to go and hence you know a lot of work we do is JSON for which document stores are ideal a lot of work we do is very deep analytics for which you need more columnar or tree based storage and hence the right storage model for the right use case is the way we approach data and we are building a you know an orchestration layer that allows us to talk to multiple data sources simultaneously which is our polygon data model cool saying and I think you know with the meetup like lots of folks here or subscribe to the open source model I don't have a lot of great projects going on the given shop so hope you guys can kind of you know learn something for the community and share something in turn and kind of we can take this whole ecosystem you know who's Scala going forward mhm absolutely thank you so much thanks for coming and you know hopefully we can host many more thanks for coming us and we're looking forward to your presentation maghreb media excellent i saw lexi