Understanding Database Technologies with Chin-Heng Hong - Episode 53

Understanding Database Technologies with Chin-Heng Hong - Episode 53

Show Video

[Music] welcome to the data leadership lessons podcast i'm your host anthony j algment data is everywhere in our businesses and it takes leadership to make the most of it we bring you the people stories and lessons to help you become a data leader our show is produced by albin business media where we make having your own video podcast as easy as joining a video call and sending an email at almond business media the stage is yours today on data leadership lessons we welcome qin hong chin is vp of product management at couchbase the modern database for enterprise applications a senior engineering executive with more than 25 years of experience chin previously held roles at hewlett packard oracle siebel systems as well as others chin welcome to the show thank you uh anthony so like we do with all our first time guests please take a moment and just tell the audience a bit more about your career before couch base and kind of how that led you to what you're doing now sure so i got my undergrad degree from uc berkeley and i stayed on for my graduate degree a master's degree at berkeley working on postgres under professor stonebreaker and before that they get a couple classes on database technology but working on postgres was the first time i was exposed to the internals of a database system and it was reprogramming list for the upper half of the system and uh this is actually the second oldest high-level programming language uh just a year younger than a fortran and i was responsible for the query execution engine and so basically using high-level language to recursively process individual notes within the query plan structure as a three hierarchy i started my first job at oracle and as part of the small team working rewriting the lower half of the oracle kernel to support the what we call transaction processing so basically we implement the row level locking and reconsistency to support higher concurrency for otp applications and if you use oracle many of you probably run into the snapshot dual errors that's uh that's uh something that worked on uh that's my first job and then continue to work on the multi-charted server in oracle seven or the object extensions for with user-defined types in order eight and then a sensibility framework inaudible as a database for the internet so basically we support the body extensions for text spatial images and other data types and let's move on to work on enterprise applications so i co-founded sarah in late 1990s to provide b2b e-commerce as a service as a manager service so at that time everyone looked to it uh down in cisco and they wanted to be just a poster child of uh doing e-commerce successfully and now our slogan is you can be adele or cisco in 90 days i've been running i will manage it for you and as we all know timing is a big factor in the success of us of a startup so we're probably too early in assess at that time and we're tackling probably too complex in application is a b2b solution and has many custom backend complex erp integrations and then the internet bubble collapsed on the two years into a into a journey so i moved on to siebel and later hp and before joining college space seven years ago uh running the product management so i could come a full circle in my career back to database that's so cool so amy you have and so this this is an episode i can already tell this is an episode where i get to like dive back into my roots as a technologist as a database person data architect and think through some of the things that i i don't spend as much time on these days and i know a lot of our audience does and so it's as i think about it you you have these different stages that you've gone through in your career but data has obviously obviously been a um consistent thread and you're working on some pretty deep technical stuff and building out features and things like oracle and now doing things with with couch base it's how do you like let's start high level how do you build new capabilities in databases and get anyone to pay attention to them because i can speak from my experience as a data architect oftentimes you're so deep in the weeds you you can barely figure out how to work with what you're currently taxed with and what your current architecture set is but it can be very difficult to think broadly about new capabilities and what you're trying to do how do you get people thinking outside of what's right in front of them i agree with you it's it's always a challenge for technologists and we have especially in the database system uh where it's so deep and broad and oracle where people spend their whole career just on transactions right so that's all they know and there's all the all they do for the entire career and the culturally very close to our customers and we start we start with customer requirements i think it's always good outside looking in understanding requirements if we look at with the benefit of uh of looking at the no sql evolution if you look at the in the early days in the in the 90s and if you look at all the different systems being developed address the shortcomings of relational system to handle at that time the new internet applications where you have data at a higher much higher volume than a back end or erp systems and data comes in different shapes and form it's no longer they don't they don't fit nicely in in the relational table and you're dealing with hundreds of thousands potentially millions of online users all at the same time globally and and uh that's not a system uh relationship that's not what religion systems are designed for what was designed for so that forced a lot of companies you know innovative companies like facebook you look at early systems from facebook netflix google and linkedin they all build internet system trying to do trying to handle this new requirements for modern applications and there's a genesis of what we call no sql or non-relational system so and uh i was exposed to it seven years ago when i went out to lunch with a good friend of mine from oracle he was working at college base i just left hp at that time looking for opportunities and he told me why don't look at my company we're doing something very interesting and we're addressing the needs of modern applications and as as i as i learned more about the company i got really intrigued by the not only the list of large customers that they have at the early stage of a young company but the type of application they're running on touch base they're all mission critical they're running a business on on the young and unproven technology and that's how desperate they are because they're running out they're hitting the walls with the religion system and uh so in the early days because we are exposed to the requirement side even though my background is coming from from a technology and that really helps that that makes it so let's talk a little bit because i think about you know nosql is still something that is people's i think it's most data architects have awareness at this point of no sequel they know they may not all have experience with it just because like we mentioned earlier there's you kind of do what you have to do and sometimes you don't get that opportunity to expand into the nosql space so it is often the second thing people think about but i'm curious how often do you encounter workloads and clients that are still using kind of traditional relational databases for things that nosql could make much better like is there a lot of this legacy relational stuff still out there that people haven't yet transitioned to a more suitable technology yeah so let me give it an example of uh of a customer that will be using cosplay for for four years now and now the hundreds of use cases running accounts may internally they instead they still using multiple databases and at a high level at a high level they they have a set of use cases they say duration is good for and understand is postgresql and then for no sql culture is the standard so i asked the architect who who chose couchspace four years ago and why so give me good looking back four years ago why did you pick couchbase and give me the top reason uh his answer is scaled down they really scale out to provide high availability right so they were running on a mainframe and mainframe is monolithic so if you have one is this high availability if there's a problem with the mainframe the whole system is down right so and also if you are going through this transformation opening up a system to the end users so sunny from instead of supporting thousands of internet users yeah something hundred thousand of the you know end user directly so the bluetooth system to scale over time uh is a key requirement that's very hard to do with uh with the mainframe where you have to keep buying big and bigger machine so those just for that company for the architect that's the primary reason for moving the couch base on no sql technology and you may have company that looked uh do notes because of the schema flexibility again let me give you another example this is the largest retailer one the largest retailer in the world and they have more than a million skills in the catalog and as you can imagine if you print out the the schema diagram of the application that can fill multiple walls it's that complex thousands of tables with all the different relationship tying them together and their challenge is business agility to introduce a new product takes weeks and months because they have to go through schema review understanding how their standard schema or whether the product fit into things existing schema if they have to change what the implication on on the existing application and that process is just way too long and the place is handled by a dba committee that doesn't understand what the applications are doing so so because of that they decided to move away from relational database and now the entire catalog service running on top of couchbase with the dynamic schema and the json document data model to just improve the ability for them to evolve the application introduce new products so so the different reasons for coming to know sql but at the high level they are very similar across the different industry so are there any i guess i i don't want to ask it out for a question but that's kind of my thing um is uh do you have any workloads or any solutions where you're like you know what couchbase isn't right you should be using a relational database uh in a different context or have you solved for using your technology anything that you would normally use a relational data definitely so so as to say to replace that existing system you have to be 10x better right else but change is difficult not just because of technology but the skill set the people that you have and the ecosystem and tooling that you build and processes around the technology that you're using today so if i'm database architect for a company i'm i'm deciding the database platform for my application my default choice today would be a relational system right because for obvious reason that we may have an existing relationship with the delegates vendor we have licenses in place but people know how to use the system we're all trained to do it and uh it's a lot easier to get get going except when you run into key bottleneck like i mentioned you need to you you you need to scale out right so if so uh i was amazed that when i talked to couchspace one of the largest technology company i'm literally the largest of this company running the most mission critic application on culture that use the profile store and they're growing they're adding millions of users every month right and uh and they at any one time they can have millions of users logging into the ecosystem and you need to authenticate and you need to authorize them you need to do that in some milliseconds that's just not what relations are designed for right so so because of that they they as a company they make they make the bet to to rely on this i then have a small company called we're talking about couch bay 1.8 it's eight years ago right so so that's to me that's a long time ago and they basically bet the company on top of it but they do in stages they they start with the culture as a cache in front of their relations system and over time then we become the system of record the main system driving all the user profile authentication so so uh there's certain things that you you just cannot do with relational i mentioned the ability to scale out to support millions of users at that scale high availability so if you have to be 24x7 and you're running a cloud environment you didn't support global deployment right you may be multiple data centers and they all need to be connected to support your global audience so if there's such applications very hard for relational to do it and if you need schema flexibility you're dealing with with semi a lot of semi-structured data you're building new application where json is your data not not just a column it's basically you know your data json they need to evolve very quickly and again so those are good reason to to do this no sql system right so it it for those folks that are a little bit less technical in the audience can you just give us a quick characterization of what what is nosql versus what is this relational database that we keep talking about uh what are the differences between these and and are there some general rules of thumb on why you would choose one over another yes uh in the early i think the they're going to the evolution so in the early days there's definitely a lot more differences no sql uh just focusing on the video scale of supporting web application focusing on schema flexibility and when not as strong in data integrity if you are doing a finance transaction you probably do not want to use a post uh no sql uh five years ago and uh so if uh if consistency if uh transaction consistency is the number one driver you probably should stick with the relations system but that's changing right so so relation system is evolving the adding skill capability if you look at the aurora look at the somewhat newer system raised to the adding skill capability even even though they're in baby steps and system no sql system i cultivated adding transactional capabilities our latest resource uh touch base seven we just added multi statement sql transaction so you can do a begin transaction you can issue a sql against your json data and they can commit and will guarantee the autonomous city the asset property of that of that transaction so definitely they're converging and over time i do expect that no sql system will take on more and more of the use cases as we as the technology mature uh we get more people trained on uh on on the new system so and let me play for you what i would typically tell someone when asked that question so typically when i think about it relational systems tend to be very good at aggregating data doing complex calculations large summations and mathematical equation driven types of um collections of information whereas nosql tends to be very good at search and retrieval types of information gaining attributes about a specific thing this is why we use it for um like login credentials and and pulling back um account information and things like that because once you get that one record then you can get the richness associated with it versus in a relational system it's very good at taking one attribute and looking at it across a bunch of different things um or or really just like using a bunch of different attributes or the same attribute a bunch of across a bunch of different accounts to do a calculation or understand sales transactions or what have you but for no sequel and this is where it gets interesting there's a lot of middle ground stuff that need to kind of do both yes and in my opinion i find that the no sequel um kind of design approach is a little bit easier to adapt to some of those middle ground use cases versus the relational side but relational in my opinion seems to always win in the end when it comes to like large-scale calculations just massive number crunching i don't know that nosql's ever going to be as good at that can would you agree or would you say anthony you don't know what you're talking about there's there's some really cool tech that we've been working on that that solves really yes i think at the high level your description the descriptions are correct for uh for most of the no sql systems out there because if you look at uh like you mentioned most of those system the early especially in the early days the focus is on performance scalability and ability to to retrieve an object a document at some millisecond response time because you're dealing with millions of users like you're releasing a a new game suddenly you have millions of users logging in at the same time you need to be able to meet that the requirement yeah and having said that like couch base our focus has always been uh marrying the enterprise grade of relational capabilities with the scalability of flexibility of nosql platform so we are the one company that take that unique approach and if you look at uh let me give you an example so we are one we're probably the only company to have a most comprehensive sql support on top of json so we extend sql to stop json rather than inventing our own proprietary language for a couple of reasons as you mentioned sql is the proven query language we're doing sim you can do simple select select style from employee web id equal to x or you can do very complex aggregation multi-way joints or union intersection and so it's a very powerful language for building application enterprise applications and we chose to extend it to support json so we can combine the benefit of both so so we see us position very strongly as you mentioned in the middle ground right when people are re-platforming their their current relations system and they need to move they're looking for a modern database and culture would be a natural platform for for them to evaluate because because then we have a family apis query language that they they don't have to retrain their engineering stuff that's i so now now now that i'm interested like well i was interested before but it's like it's it's it's thinking about how you're able to start to blend those and this is something that i've seen manifest itself in you know a database engine space i've seen it in them in the metadata and data catalog space it's it's like you have these kind of general um you know functionality that gets associated with a particular kind of technology and then all of these vendors start to spring up that start to blur the lines between everything and so you have to understand like okay what's our textbook definition and then why is vendor abc's offering kind of like a dash of this and a bucket of this and then a splash of this thing over here and it and it's really interesting but if you don't understand for the people out there uh that are listening to this kind of like i don't understand half of these words it's there's you have to understand some of the fundamentals before you can really understand what a vendor's approach to what they're doing really is because they'll tend to exist in a place that blends these things in a unique way that may very well suit your particular needs exceptionally well because they're generally doing this because they've identified a need in the marketplace if we had these technology capabilities all in one and could solve for some of the deficiencies of historical products in this area hey we gotta get business here let me give you another example so so very well say so if we look at look at smartphone right so smartphone is a combination of we use carrier phone we have emails we have a page with email system we have a we have a music player we have a gps navigator now they all combined into a single platform the iphone and it's just not the summation of the part it's the integration that allows you to build very interesting application and and give and take the user experience to the next level so for culture we're doing the same thing we're integrating multiple services in into a single platform you mentioned about very highly scalable key value access that's where our foundation is we had a document database reflective schema we we support poorly indexing so that the retain declarability and doing complex uh operation as as you just described and we with analytics so so so we describe ourselves as a hybrid transaction operational operational and analytical system so if you look at a lot of modern application in real time you need to analyze a lot of data and use that data to drive personalized experience and we actually have a analytic service that can run independently of the operational system in one single cluster so that your one single system and you can do complex we actually have a mpp engine that can run very complex query in real time and do a lot of summation aggregation and find the data and use insight to drive the interaction all within a single system very cool so and that's it helps to think about things like the phone right the phone is something that we probably have most of us have one within arm's reach right now wherever we are we might be listening to this show on a phone it's quite likely and it's that integration and that blurring of lines and that that reduction to what's what is necessary is it simplifies the user experience and it does so with enormous complexity behind the scenes and that's correct that's correct i often think about how um you know being simple and being concise is a extremely complicated endeavor and it and it takes a lot of effort to like write a short email you know that classic apology you know i would have i would have written less but i didn't i didn't have the time to do so you know this is especially true for a distributed system this resistance is very complex so as you say our goal is to make the complex system very simple both for developer architect and also for the for the devops person that's why that uh if you look at couchbase uh whether running on the desktop or we are running on the hundred node cluster they're all the same we auto partition the data we we automatically replicate the data for high variability if the nodes go down we also recover and fail over and rebalance the data for you all without interruption to your application so making complex things simple it's very hard and but uh but it's very beneficial to to your end users yeah it and it's true it's true in the product and technology space it's true in the consulting space one of my that's correct i'm coaching consultants i tell them your job is to make your clients life simpler to make that easier don't bring the problem to your client bring the answer bring the solution to your client and that's where you're going to add value and so you know technology does the same thing is it's nobody really cares about the technology they care about what the technology does for them and then if you can if you can connect them to something valuable then your technology is valuable too yes and phone is another uh good way to understand the new types of application that enterprises are building uh we use i use our phone to do all things today right so to make a reservation i just actually came back from doctor's office instead of filling out a form i actually registered online because the easy i see uh registration so that i can walk in and see the doctor right away and if you think about i'm a ds is one of the largest customers they're running 20 million ops per second on top of couch base at that scale and think about i think we're old enough to to remember when we used to make a travel plan we call up a travel agent right say i'm going i'm going back in this case i'm from malaysia i'm going back to malaysia can you help me look at what the flight availability between uh during christmas so come back with with a few choices we either over the phone a few times then you book your ticket but nowadays we all go to kayak or expedia and we do a slicing and dicing look at all type of options and and we're using a phone or a web browser and the requirements application trends change drastically and in the old time the travel agent log into a mainframe or a large database and probably hunt tens of thousands of them doing that at any one time with the port like kayak or expedia now you expose it to millions of users and they're all doing all kinds of research or searches or are using a phone or on their computer what we call the look to bulk ratio so that's the the term that they used to describe the number interaction that you do with the system before you you perform a transaction changes from a few hundred to one to in my ds case three hundred thousand to one it's basically a thousand foot increase so we're all doing all kinds of requests before we book a ticket right and you need a system to be able to handle that increase in interaction with the end user to provide that personalized experience and before they they do a transaction so be able to do that efficiently at a scale that you need globally require a modern database platform that you just cannot do on your legacy as a system right right so i want to spend before we start to run out of time because i could talk to you about database technology as a couch base this entire time um but there's a you guys did a um kind of survey and a research study around digital transformation and digital digital architects and and we want to talk about some of the the results from that because i think they're very interesting especially in the context of the pandemic and all that so you can talk a little bit about what that effort was about and what those those findings were and we'll talk about some of the details beyond sure thank you we basically surveyed about 450 architects in united states united kingdom germany and france or the country that we focus on and so that's a bunch of questions about the challenges that they have and what they're doing with technology in the digital projects and a couple of findings are pretty interesting first of all as you know uh digital transformation has been the top of cio agenda for for many years now and with the pandemic actually put a lot of pressure on on the enterprises on the businesses to both change and update the existing system uh to handle new environment and also to deal with the all the new remote working environment how the team collaborate and we half of architect actually responded saying that the under extreme pressure high pressure to deliver on the digital projects today or last year versus a year ago which is about 19 so there's a three-fold increase in number of people feeling the pressure because of the whole pandemic uh a situation and adopting technology is a key part of making sure that your the transformation uh is successful and so we ask questions on what technology are you using that helps and doesn't help and not surprisingly cloud and big data are the two technologies that uh more than half of the architects say it will be they believe will be transformative in how they're doing the new projects and the interesting thing is many of them are still using relational databases as you just mentioned but most of about 61 of them say that relations is not holding them back for the reason that that i just described when you're opening up your system to to the online user the the type of requirements are very different than than what you're used to uh dealing with so they're all struggling with we're looking at evaluating new modern database and no secret is an obvious choice and we believe that by providing by and skill set is always a challenge right moving from a system that you know well do a new system understand the architecture understand the api the query language so we try to make a stimulus for for enterprises and by providing the sql interface with json that really helps them and also help them move in stages right you don't have to rip up the whole system and replace with new one you can do that over time like i mentioned the largest customer that we have eight years ago when they adopted cloud space is to use it as persistent cache in front of the mainframe and then over time they will become the source of tools they integrate data from up just multiple system and eventually they they actually send a picture that they roll off the truck the mainframe of the truck right so but they took them it's a multi-year journey and uh and we believe that we need to help a customer go through that transition in moving from a legacy to a new technology yeah so i i can relay as a as a digital architect myself you know some of the challenges that that i saw in in with the covet 19 pandemic and kind of a quick and unplanned shift to a widely dispersed workforce and fully remote and all of that um and there were some interesting dynamics that i and that i wonder um you know how it impacts what we did and are doing now and also i i find myself think you know being so happy that at least we had the technology to continue working the way we had because i imagine like if we didn't have the video chat functionality and the cloud backbone for so much of what we're doing with data and technology today things would have been much more difficult to manage remotely um yeah can you talk about some of the findings in that that survey around that like with the cloud and with some of the remotes you know the shift to remote work what changed in terms of the architecture on the on the back end did we have to do anything really quick to put out fires or were we ready for this and we maybe didn't even realize we were ready for this yes if we just look at uh as a consumers our daily lives and how we interact with the with enterprise system like i mentioned i just i just have a doctor appointment i log in i register online going to restaurant now you scan the qr code you're no longer looking at physical menu and so industry of the industry the experience is moving to to digital and and the pandemic just accelerate and put more pressure on company uh to do that and so you need a system that can provide personal experience for on mobile and and web and union system that can deal with uh deal with a lot of personalized data right different different form and shapes and uh and dealing with the increase in interaction right people looking at manual now they look at they want to look at photos they will look at reviews that what my wife does all the time go to the restaurant and just look at all the pictures and figure out which with dish to order any longer looking at the paper menu and they put a lot of tremendous load on your system that you do you do not have with the prior application so you need to think through that that new requirement anything to how how do i architect my my system to handle that new requirements and how do i do that in phases right it's very hard to rip and replace the compact system overnight and having a system that allows you to phase your changes in is is critical to to the success of digital projects yeah that's a that's a great point for non-technical folks as well as because we we think about change management and that that people and customers especially they can only handle so much pace of change there's only so many different things you can ask people to do all at once so you have to think about how can we ease people into a new way of interacting with our organization and that's true on the customer side and i think that we can think about things like menus and and how people order food and and do things that we do all the time but then also on the employee side as we serve our employee base and and help them help us right how do we serve our employees so that we can put them in a position to be successful whether they are remote working when they're when they're in the office when they're traveling for whatever purposes that they have how can we give them an experience in their work so that they can do their best work and so that they can help our organization grow in the way our organization needs to to serve those customers who we are now engaging with in ways that for many organizations they hadn't done before the last year or two that's correct yes so i guess we're we're lucky to live in an era that we have a lot of tooling help to help us work remotely the video conferencing is one so so we're now using a lot more frequently than than uh than we used to a year and a half and and a half ago the ability to do messaging instantly with with groups and one another that helps the communication then you definitely need to find a way to bring people together whether it is informal happy hours virtual happy hours that we do once in a while just just to make sure that we continue to have the informed interaction we're missing a lot of the of the hallways conversation and knowledge transfer in the remote environment so we need to find new ways to keep in touch and it's challenging for for uh for a lot of companies and many of us are used to working remotely so it's just a it's an easier transition for some of some of us than others but surprisingly in a survey architects and most companies are responding quite well to the challenge and more than despite the increased pressure to keep the lights on for assisting application and accelerate new projects and about 40 of the architects just say they are delivering on that digital project and interrupted by the pandemic with that that is a good sign but it is it was still a troubling side one of those those uh figures that was were in that report is that level of pressure and stress was definitely more pronounced now than it had been in the past definitely yes and i and i think you know when i think about it it's you know just like we have this convergence of the human and the digital in terms of how the customers are engaging with our businesses i also see and i've seen this for a long time this is probably decades in the making is the convergence of the technology application and the technology data it used to be in our organizations we had a data group that was doing data and back office stuff and building our our databases and then we had the application group that was working on user interfaces and yeah i would venture to guess you've written a click event or two in your life as well and it's like these used to be totally parallel groups and have converged greatly and i think kind of nosql sits at the heart of a lot of that yes i mentioned that uh h higher with high velocity scale is one key requirements the data flexibility and being being able for an application group to be responsible for your own schema and the evolution of it is key because with the new macro service architecture you have a huge application is not broken down into micro services and their different scrum teams are responsible for each micro service and their response for their own schema and in many cases the choice of the database as well so putting the response back into the application a lot more agility and enable the ability to really evolve uh quickly and and we all and nowadays if companies are rolling our application on a daily basis changes application and you're just going to do that when you you have mores between the groups right then then you have to go to committee to get a change approved before you can implement something yeah so from your perspective then as a person who works for an organization who's building you know this this technology and at that forefront of data and applications and that convergence in the cloud and all of these scalability uh considerations how do you recommend people go about managing managing a technical career for like those folks that are looking and saying wow i've been doing you know data warehousing for the last 20 years or maybe a college student that's getting ready to graduate and says you know i got to figure out where do i want to focus what kind of advice do you have for about what what's coming up like how do people manage their careers in this time of you know confusion and uncertainty and convergence and all of these things that are kind of mashing together to create these experiences how do you pick a spot or like what do you focus on to become effective in your career in this current time yes being curious is good so as you mentioned things are changing so rapidly and we constantly need to relearn and a new technology but the good thing is for myself personally in the data space conceptually things doesn't change that much transaction the concept transaction is still valid people should talk about acid right this is implementing that in the new environment new architecture is is is challenging and in some cases interesting right so so uh taking what we used to do in the with the modernity system and applying that in the in the world where things are distributed uh and eventually from we're going to push all the way to now we're looking at as a company looking edge computing a lot of companies happening at the computing is happening at the edge and a lot of interaction happening at the edge and how tall do you when you have from a few data centers now you're exploding to hundreds and thousands of edges right and how do you manage that complex environment so and uh being curious continue to to learn for myself is understand the the space that i'm in and tracking the the advances in their space the good thing is that they're so much on that you can you can research on the web and if you want to get your hands on the system there's so so many systems that provide free online account that they can sign on that uh they can just sign on and and start doing some simple prototyping just lend up the technology so so i think time is a challenge so if you have the curiosity if if you're willing to learn i think there are lots of uh channels that that allowed you to uh to do that yeah i would and i would add to that my own advice is you know understand the fundamentals understand when we say sql understand how to write a query and understand how some basic relational data does work you don't have to become a nest you know an expert necessarily but dabble with it understand a little bit about it understand what documents are and key value pairs and how to do some basic json texts and and just understand what that world is about and then understand how programming works even just pick up some python python is accessible you can you can understand that so get some of these kind of basics in the core areas and then start to understand okay how do these different technologies twist them how do they bring different strengths together in unique ways and then make it so that you can scale greatly or that you can have acid types of transactions in a nosql environment or how like those things start to make sense if you have a basis of the foundational principles and then to your point i think is as strong a point as can be made is be curious unders like try to understand how this stuff works and let yourself meander through a path under like figure things out and i think that's a good advice as well because you have to understand that story you have to understand that connectivity it's not all just about features and speed and and and attributes there's there's a story behind why these technologies exist and that's why i like to have these conversations on data leadership lessons is that we get to learn a little bit of the thinking behind something like couch base where like we can go out on the web and learn hey this is what couch base does and it's good at this and the reviews say this and this is what the cost model is all that stuff is fine but to understand some of the thinking behind the product from a vp of product that to me is why we do this show is so that we can understand some of that story behind it and connect it to what we might be doing in our career or what our businesses might need i agree with you because at a high level they all look the same improving tco improving user experience performance and skill but you need to double click on on exactly what sweet spot and what type of use cases they're addressing and how and why the assistant system do not address those well yeah and so for those folks that are thinking about couch base or thinking about other technologies do you have any recommendations for like how do you evaluate these tools when you're to your point like they all kind of sound the same especially when you're not an expert in this area like i've got a team i got to do data stuff i got to figure out a technology selection and how do i start to approach that in a way that that makes sense like how how do you recommend people do that so uh the a lot of open source product culture has one is one as well and then most all of them if not most of them are all offer of community edition so it can get a hands-on on the software easily and you can like say that the cloud service you can get a free account easily just uh like you say get your hands dirty and just start prototyping think about an application that you want to build and uh like with our mobile product one of the first examples that we show people is a to-do list how to showcase offline capability when when the network is there you can still edit your to-do list and then when the network's back you can you can see it's synchronized with the cloud database so do something a simple application many times will help you understand uh the technology and and what they're addressing uh uh behind that so so get getting hands-on learn about the technology uh but build some simple application on your own and i think that's a good way to start i think that's great advice and just like that we're out of time like that went by so quickly i was actually astounded i'm like oh my goodness we've got over 40 minutes already um but i think that's a great note to end on is is how to go about taking that complexity that is surrounding us and starting to make it simpler and more actionable and in a way that um helps us solve those problems that are like like your research said are more pressure-laden and more important to the success of our organizations than than they ever have been yes and uh i know it's it's been pleasure talking to you it goes by a lot quicker than expected and i was actually nervous that i don't have enough topics to the last 45 minutes but it went by so quickly well thank you so much for for being on the show and and i've done my job if i can get you to forget that you've been on a podcast that we're having a conversation for an audience and then i'm doing my job so i'm glad that that went by quickly and i really enjoyed our conversation thank you thank you today you too and thank you all for joining us today you'll find more information and links in our show notes dive deeper with my book at dataleadershipbook.com and use promo code augmentdl at the dataversity training center for 20 off your first purchase please remember to follow data leadership lessons on youtube or wherever you get your podcasts and if you enjoy the show please rate and review and help others find us stay safe during these unusual times and go make an impact [Music]

2021-10-12 08:59

Show Video

Other news