ASF 008: Michael Rys interview

Introduction

Michael has been doing data processing and query languages since the 1980s. Among other things he has been representing Microsoft on the XQuery and SQL design committees and has taken SQL Server beyond relational with XML, Geospatial and Semantic Search. Currently he is working on Big Data query languages such as SCOPE and U-SQL when he is not enjoying time with his family under water, on the ski slopes, or at autocross.

This talk has taken place during PASS Summit in Seattle, WA, on 2nd November 2017 (Thursday).

Do you ever wonder how many people work on the USQL or Azure Data Lake? What animals does Michael see from his window?
Listen to this podcast to get to know what two career choices Michael had had when he finished PhD in Switzerland.

Transcript

Kamil Nowinski: Hi Michael.

Michael Rys: Hi.

KN:  Thank you for accepting my invitation for that podcast. Basically this podcast is not only about the technical stuff. It is also about to show you as a person, your habits and other things, what are you doing on a daily basis, etc. So let’s start and could you introduce yourself to the people?

MR:  Sure, so my name is Michael Rys and I am a Program Manager at Microsoft currently working on Azure Data Lake and our internal Big Data Platform called Cosmos. Not to be confused with Cosmos DB. I have been with Microsoft for almost 19 years, started in December 1998 originally working on XML support in SQL Server and then I have worked on SQL Server stuff until about 2011-2012 doing the beyond relational scenario in SQL Server 2012 when we shipped things like FileTable, Semantic Search, more spatial support, xml indexing and couple other things. Then I moved to the Big Data space back then.

KN:  Where are you living exactly?

MR:  I currently live in a small suburb of Seattle and Bellevue called Sammamish which is about 20 minutes’ drive from the Redmond campus…

KN:  I know where it is. My friend lives there.

MR:  Ok, cool.

KN:  Because you mentioned you are from Germany…

MR:  No, I am from Switzerland.

KN:  Oh, Switzerland. I was thinking for a while that you are living and working remotely…

MR:  No…

KN:  So you are a local guy here…

MR:  Yes, I am local, living here for 19 years.

KN:  From which version of SQL Server you are working with?

MR:  So I basically joined the SQL Server team for SQL Server 2000. So I was working on the first set of the XML features that we added in SQL Server 2000.

KN:  Good to know… So you were behind the scenes…

MR:  So you can blame me for FOR XML and OPENXML and some other stuff.

KN:  Good to know and if we had some question not only about the U-SQL so we can ask you about the XML and other stuff.

MR:  Yes.

Damian Widera:  But Michael you are not involved in the JSON stuff which works in SQL Server

MR:  So the JSON stuff in the SQL Server – I did talk to the people when we started adding it to give them a little bit of background information about the design decisions that we made with XML because JSON is very similar to XML in the terms its hierarchical format. I wanted to kind of give them a little bit of a hint to make it similar so people that know how to process one hierarchical format can very easily apply the knowledge to operate on the other hierarchical format as well. But I was not directly involved. That was actually done by our team in Serbia.

DW:  Ok, but I can see the consistency.

DW:  So it is really easy. If you have any experience in playing with XML in SQL Server no matter in which version than it is easy to adopt the previous knowledge to work with JSON. So that’s fine! We don’t have to learn new stuff.

MR:  That’s kind of the idea that was around that

DW:  I would like to ask you – do you have any favorite set of specific tools you are using during your daily work? I mean when you have to play with SQL Server or USQL or Data Lakes?

MR:  You mean besides Outlook?

DW:  Yes, of course.

MR:  Well, as a Program Manager of course some of the things that I do a lot is one things is helping customers with their questions about how to do certain things, how to debug certain things, how to optimize certain things. And also obviously preparing demos for both blog posts, documentation or presentations. And for USQL right now my favorite tool for doing that is basically the IDE extension in Visual Studio. So that’s where I would spend most of my time in like looking in job graphs trying to figure out what people are doing, to help them improve their query performance or improve their query efficiency and of course their cost efficiency. So some of the things I would look for example is the vertex execution view and the modeler are the two things I would often use to help people improve the cost efficiency and understand the jobs. Another thing that I obviously do when I do my own development is basically to use the Visual Studio development experience for doing the USQL and code development. Otherwise, one thing, one tool that I would like to get to know better is our VS Code tooling around Data Lake. So I am kind of planning of probably doing one of my next demo project is to use just VS Code instead of Visual Studio for doing that.

KN:  I noticed that there is an extension for USQL.

DW:  Yes, there is.

MR:  So these are the things that I am currently using and obviously the other things are like a Notepad++ for looking at results and understanding encoding and documents etc. Otherwise it is just the standard program management tools like Skype and Outlook and One Note and Excel and things like that.

KN:  and you are working with USQL and Cosmos and what do you think what is the current direction of all those stuff?

MR:  Well, one thing you notice is that there are a few kind of trends that are occurring not just within our offerings but also in the competing offerings. You will see that people are getting away from the low-level clustering management and low-level programming paradigms and they go into something we call “serverless PaaS offerings” like Azure Data Lake Analytics is and the ability to operate at the higher levels of abstraction. So the serverless PaaS offerings is to manage the costs better. Like you do not have to manage clusters, you don’t have to pay for provisioned things that you are not using. You basically have shared resources that you have if you need them. But if you don’t need them they are basically available for other people to use. It becomes like a self-managed backend system. Another thing that I see is that the higher level of abstraction in terms of programming goes into both going towards SQL-like languages for doing the processing like you saw that with MapReduce when it went to Hive. You see that with Spark when it’s going from the LINQ-like extension for Scala and Python and Java towards like-Spark-SQL. You see it happening also at the level of what people actually would like to do with it the data. It is going away from just doing SQL like analytics into doing Machine Learning, into doing more domain specific processing of the data because the data becomes more unstructured when you operate in these environments. So want to have more domain specific code that is scaled out. So in Spark you see it with the Spark ML having that. In USQL you see it with our extension libraries with the whole framework extension …

KN:  Cognitive Services…

MR:  Yes, Cognitive Services in this aspect and so on. But also the framework we provide to people who do it themselves.

DW:  Yes, I am thinking about a little bit different thing, because I would like to ask you what role do the MVP play in the – form your perspective – how MVPs can improve for example the technology you are responsible for?

MR:  So, I mean the MVPs have always been very instrumental in couple of ways in the technologies I have been involved in SQL and now in the Azure Data Lake and USQL. I see them as one early adopters. I mean I expect MVPs to often be like people that are using our things at the beginning, give us feedback and basically are the first external customers so to speak like to give us feedback and then use it. Also potentially contribute things to it like especially with the extensibility module that we have – we have some MVPs that provided let’s say an Excel extractor. Tillman did that recently.

DW:  Yes.

MR:  So that kind is one role. Another role is of course around amplifying the messaging. Like go and teach because if it just be us teaching then it would be fairly… as we would not scale as well. So having MVPs helps us with the teaching and bring their own standpoint. And I bring the feedback that I hear from them is another way. And another way is to build community engagement by running user groups not just for our specific products but for the whole data platform aspects as well. And also help customers to understand how these different components that we have in our ecosystem in the Azure fit together and what are the sweet spots to do so. And last but not least obviously it is also basically helping amplifying our messaging for example by inviting me into the podcast.

DW:  That’s right.

KN:  Ok Michael, tell us something about you. Tell us something about your hobby please. Do you have any hobby?

MR:  Yes, I have a few hobbies like you were in the presentation today and you noticed that it was a green car on the race track. So that is one of my hobbies to take my green car into the auto courses and try to improve my driving skills. Another one is scuba diving although I have not got too much time recently in the last years to do that and in the winter I like to go skiing and spend time in the mountains. That’s few of the major hobbies that I have. If I am outside of writing about USQL scripts and so on.

KN:  And very popular topic that is recently – work life balance.

MR:  Yes.

KN:  What do you think about it? How you can manage that?

MR:  Yes, that’s always a hard question. The more fun you have at work the more work life balance you have in term of yourself. But if you have a family than you have to make sure that there is time for family as well. Let’s say this way. At Microsoft at least, there is an understanding that people should have a work life balance so there is a little bit more agreement around that that we might be out of the weekend and not necessarily available online except for if you are on the call rotation. Which is a little bit more work life balance than let’s say if you are in the startup or so. But it still is a, I would say when compare to normal job, if you are a passionate about your work then there is more work…

KN:  at home…

MR:  at home and there is more work as part of your life.

KN:  Yes, so the passion

MR:  so you always have to like a trade off a little bit.

KN:  But it’s also hard to explain this to our wives.

MR:  Yes, but I think that yes, you have to take time off at some point. Either by taking some vacation or reserving let’s say weekend to like spend time with the family and so on. But yes, if you would talk to my wife she would be probably think than I am first married to work… But from the other hand we are still married so…

KN:  So it works. It is good. But you are living very close to Redmond. It’s about 15 minutes by car.

MR:  It’s about… let’s say – it’s about 20-25 minutes.

KN:  Depends on the time

MR:  And depends if there are police officers on the way

DW:  But you have a sporty car so it should not be a problem.

MR:  Yes, the car is not the problem.

DW:  That’s right. Ok, so maybe a question like this. What hints would you give to young people who wanted to start working on the market?

MR:  You mean to start working…

KN:  on the IT market

DW:  as the IT guys…

MR:  Yes, that’s an interesting thing. The first thing I would say is work on understanding your customers. Like regardless there are a lot of interesting areas in IT but I think that the important part is to be successful in the understanding what problems do your customer try to solve. And then understanding what are the best tools that help you solve it. And obviously you wouldn’t have to have some foundation in terms of understanding the technology and the technologies base and potentially different offerings in that space like maybe become an expert in Azure and Azure Data Lake – that’s what I would say. But also Spark definitely would be something if you want to be in the Big Data space. And then find basically the vertical customer segment that you feel you can understand well and be passionate about and kind of provide solution in. That’s if you want to be on the like IT side. The you can bring that into ether as becoming an employee of a company in that space or becoming a contractor or consultant in that space. If you want to go into actual building the system then I think what helps again is understanding the customers etc. but also understanding the technology. And we have MVPs that ending up being a part of the product team. For example not just because they are passionate about the technology but also contributing to the technology and the go through the interview process and basically understand how to be successful there. So it is all, I think the most important part is that you understand in what area are you interested and passionate about and what areas do you think the customers can benefit from your understanding and passion and capabilities.

DW:  so you said a very important thing because apart from the technical knowledge you have to have some communication – how to communicate to customer, how to solve their problems and this is something that can be very tough for young people. They have to learn a lot I think from the older colleagues to have the ability…

MR:  I would not disagree with that, yes, there are obviously ways of how like they can learn that like they can engage in the community events and to speak or just to attend meetups and like to listen to people talking about the problems and so on. Yes, it is more than just building code. I mean when I actually went into the IT area when I was 19 years of age I learned programming at high school. I kind of liked it but then I started using it as a making some summer pocket money. I worked for some local companies and the university as a programmer. And you quickly learn that your programming skills are need definitely to how solve the problems. But understanding the problem you actually need to code for, understanding the customer demands, being able to communicate with the customer to translate his “do what I mean” into actually telling him what he means and being able to actually build the solution is almost as important if you really want to be more than just, what I call, a coder. I mean you can find the niche where you basically code what somebody else tells you to code but I am not sure that something that is fully fulfilling at that point.

KN:  Yes, but this is ok when you are working as a developer and you are very close to the customer like in that case you are describing. But in the biggest company or let’s say – the corporation – nowadays. I think that the developers are doing separate things and BAs (Business Analysts) doing separate things. BA is contacting with the customers.

MR:  Yes, but even we at Microsoft, like me as a Program Manager, so I am not writing the code by myself. I am translating the customer demands into design spec and then the engineering team can implement. But even in the engineering team the engineers are normally providing better solutions if they really understand the customer problem and can communicate back to me and others about what they want to build so we can assess if the solution is actually going to help us address the problem or not.

KN:  Yes

MR:  So I think the communication is still important.

KN:  Yes, communication is the key. And you as a Program Manager are on the “dark side” currently

MR:  Yes.

KN:  I have heard that developers called that.

MR:  But we have the cookies so that is nice.

KN:  Do you have some animals at home?

MR:  Not at the moment. We had some rabbits…

KN:  Oh, rabbits…?

MR:  yes, and chickens, but they were all eaten by the wild animals around us. We have eagles on one of the trees next to our house and we have visits of bears and coyotes and our neighbors sometimes see cougars and lynx.

KN:  So all those animals are in your area?

MR:  Yes, and some of the neighbors have dogs so we can sometimes use them as pets so to speak. So right now not. Chickens got eaten, rabbits got eaten so at that point we decided to give up.

DW:  But this is an unusual answer. Typically we hear about dogs and cats and sometimes about parrots …

KN:  … or fish …

DW:  … or hamsters but I have never heard about coyotes or bears.

MR:  I would not call them pets but they are living in our neighborhood.

DW:  My next question would be: how are you preparing when you have a conference like this or you giving lectures to user groups. How do you prepare yourself to do a speech?

MR:  Well, I have been giving presentations now for a long time so the preparation is prematurely along the following mentions. First if it is a new topic I obviously have to build up the material, understand the story I am going to tell and make sure that basically the slides and demos kind of help me tell that story. If I have given the presentation a few times at that point in time the main focus is to make sure that the demos run and that I am not having any surprises with let’s say the flow of the slides. Sometimes the PowerPoint seems to be losing my animation orders and stuff like that. But prematurely it’s like understanding it again, understanding the audience, what is the level of the audience, what is their interest and then tailoring even the previously given presentation. Tailoring it to that audience either by making the story telling more relevant to them or adding some demos that is more relevant to them. But one of the secrets of successful presentation is normally that you have a set of base stories that you can kind of combine into different things and then have different levels of your story telling. So if you talk to CEO type person you focus more on the business value on like understanding the high level picture versus if I am going and talk to the audience at SQL PASS then I am probable more talking about like coding experiences. Or if I go really deep I go and talk about like internal architecture so that people can better tune or performance tune or debug their performance of the product. So it all depends of the target audience. But the main thing is to understand the story you want to tell, make sure that your slides go logically into that story, bring people along on the story so that means you start a little bit high level and then you dive in and make it relevant to that audience. Like if you talk to business people – make it more business focus. If I talk to my field people then I need to sell that and make it more relevant for them so they know how to sell it. But if it’s for engineers make it relevant to the engineers so they can understand how to use it, how to optimize the performance without the debug and so on.

KN:  Do you still feel nervous before the presentation?

MR:  That depends on how well I slept the night before. Normally I am not felling like this anymore like I used to when I was like in early days of my presentation career. Unless it’s like in front of thousands of people. So the audience size matters a bit in terms of nervousness or if it’s like a keynote presentation where you are like supposed to give a keynote that’s probably more nervous than a topic that I understand deeply. It also depends like if I own the area of the or am very familiar with the area then I am less nervous than if it’s something that I have not presented to the audience or if it’s like not necessary my area of expertise.

KN:  Yes, makes sense.

DW:  I would like to go a little bit more to your background. When did you find SQL Server for the first time? Was it during your studies?

MR:  So I did do my PHD back in Zurich at ETH or Swiss Federal Institute of Technology for those that want to know the English name and during that time my professor Hans Schek, he was taking classes at Microsoft research back in the mid 90’s and he came back and talked to us about the OLEDB and SQL Server being re-architected using OLEDB to like make this composable infrastructure architecture in SQL. In which if you look at the SQL Server you probably would not notice if you looked from outside but internally it’s still like it uses for example the rowset provider like the connection between the storage engine and compute part of the engine. So that was kind of the first time that I have heard about it. The first time that I really got in contact with was when I was in Stanford doing my post doc and SQL Server was just launching SQL Server 7.0 and the research group led by Jim Grey and Microsoft in San Francisco basically had an event that I went to and I saw the SQL Server 7.0 for the first time. It had some interesting capabilities like linked servers and English query and I found it kind of interesting. Although it wasn’t really the reason I end up working for Microsoft I must say. At that time like I was building my own database systems in semi-structured data management project at Stanford. I was working on a project and before that I built up my own scale out update SQL database or object-oriented database on top of Oracle. Now one thing I knew was when I finished my PHD in Switzerland I had two career choices. I could become an Oracle consultant and make a lot of money helping people understanding the complexity of Oracle or I could go and build a better database system that could make the consultancy superfluous. So I basically decided to go and build something better than the complexity that was Oracle 7 at the time.
Did that answer your question, Damian?

DW:  Yes, yes, of course. The next question would be, because you know we are, me and Kamil and other people, we run a community which is quite big in Poland and how would you convince people to be a community leader or at least a community speaker? Because we have a lot of great people but they are somehow afraid of talking in public.

MR:  Oh, well, in the thing that I think the community engagement helps you with is getting over that fear. I mean you are among people that are basically your peers. You have people that know the area, that seen you more as a friend. It’s not like in work where if you give a bad presentation you probably have a few people that will be cheering that it was bad, so then might get a promotion not you but in the community it’s also all about helping each other. So going to the community and then speak gives the ability to exercise your speaking muscles so to speak. Learn from it and basically get the confidence to then speak on eventually more important events as well. Also being in the community, I think, you often like one of the things that I feel the more I am helping on the community just being online in the community like, I don’t know, StackOverflow and events like SQL PAAS or the event where I was speaking last week with you guys. I feel the more I talk and contribute actually the more I learn. And the more I think I grow personally as well because people have feedback that is potentially unexpected could help me understand where they are coming from and applying that in my work as well. So I think contributing, participating actively is overall net positive. People are nice, they are encouraging you, they are here to help you with the feedback and so on. It’s not like they are there to kill you if you are pronouncing a word wrong or have to many pauses or ‘s in your speak. It’s all about basically making each other great, helping each other out and everybody has something that they can teach somebody else.

KN:  So they actually shouldn’t feel any fear.

MR:  No, they shouldn’t be afraid of that. I mean they can potentially just talk about cool code they have written and the thing that often helps is like if you present on something that you are really passion about then the passion shows through. I mean, if your demo is crashing or so on, people will forgive you that if they notice that you are really passionate about the thing. And you will learn. I mean it’s all learning experience. Nobody who is presenting on at the high level did that first time he presented. So I remember the first time in the high school I presented. I wanted to go to the floor and hide. If you see me now presenting you wouldn’t think that has ever happened before.

KN:  Tell us, are you perfectionist? In your work, in your private life?

MR:  I wouldn’t call myself to have OCD but I am definitely trying to have a like optimize the perfection to the level where you can deliver it. I mean obviously you have to sometimes deliver – deliver is a feature is what I normally say and which I learned from some of my early mentors at Microsoft. So being upmost perfectionist and never deliver is obviously wrong thing to do. However, I try to design the things in a way that makes it easy for people to use and kind of have a nice experience around it. I mentioned I drive car for fun and so one of the things that I notice is that some car manufactures are better at like optimizing the benefit cost tradeoff between giving you a good driving experience and still make the car affordable then others are.

And you can kind of see if you have this then you are to some extent a perfectionist. But you can learn how to deliver something that is maybe not a 100% perfect in the first release but allows you to improve perfection up to the point where the return of investment becomes negligible. As an example, when we did the XML support in SQL Server 2005 one of the things I was never happy about was the performance and the size complexity of the XML indexing that we had at the time. It was good for certain use cases, it was very good for some like a generic use cases but it had a little bit of a high complexity in terms of optimization and it very quickly led to very huge indices and so on. I was kind of my perfectionistic things would have wanted to have better indexing story right from that but we needed to shape the things. So we basically left what we had and then we improved on it based on the customers feedback and learnings that we have. And we had some really like a helpful customer and interactions and then this helped us to deliver the XML indexing feature in 2012 that gave you the path based indexing and so on which is giving you kind of better experience. Which is a way much better experience at that moment of time. So sometimes you have to shape and then would I have done more in the later things – yes, but at some point in time you reach the point where you basically polishing something that is already looking perfect for most everybody else. Then you probably should stop and do something else.

DW:  It’s not our first interview where we can hear that the most important thing is to love to learn and love to share the knowledge. And it’s great that we can still hear the same from you that you have to learn a lot I think during your career. You have to be very active and sharing the knowledge is also very important I think because from that point of view you can also learn from the other people. And this is what we are doing all the time. But my last question would be – which achievement are you satisfied the most? I think you have a lot of them but what is the best achievement for you during your career?

MR:  You mean work wise? Work wise… I think the achievements that I am most proud of is when I basically go out with some capability and I see people using it and providing both their pleasure and sometimes the displeasure about the feature and how it helps them solve their product. I think that’s what I kind of feel is my, what I am proud of. Basically, that I have to deliver a lot of things to both SQL Server and now in the Azure Data Lake and so on. That really helps people solve their problems. That was the reason I didn’t go to the academia but decided to go to like a development and research to help solve peoples’ problems. So if I see people using my stuff to solve interesting problems that kind of gives me huge satisfaction.

KN:  I like the sentence you used during your presentation when you were talking about the version 1 of the Data Lake that people “love to hate”.

MR:  That was about Azure Data Factory you mean?

KN:  Yes, about the Azure Data Factory. I noticed that you are not a fan but you very frequently use the Twitter.

MR:  Yes.

KN:  Are you a fan of Twitter? Why do you use that way of communication?

MR:  So I mean I use Twitter prematurely as a way of getting information out. It’s very like a low friction way of communicating with a large amount of people and allows people then re-share it easily. Am I a fan of Twitter? For that purpose – yes. There are some other purposes, especially in the US right now that I think Twitter is a little bit over represented and so on. There is all like the Twitter bullying aspect and so on that has fortunately not reached out our area as much as some others. I think it’s just one of the tools that we have to get your message out and keep engagement, two ways engagements with the community and with people that you want to have interaction with. It definitely has been quite useful for sharing information very quickly. Also listening to what people are talking about in the areas that I am interested in currently. Basically, if you use Azure Data Lake, USQL with or without the dash in your tweets I will probably pick it up to see what are you saying and then if I have something to contribute then I will reply to it.

KN:  OK, and the really last question is about the team you are working with. How big the team is? And how many teams are?

MR:  So, I actually don’t know the actual number right now but basically on the one hand we have the PM team which is a not just me but my management and my colleagues that cover Azure Data Lake Storage and Analytics, different aspects, jobs, code experience in Visual Studio, development experience and the actual language experience. Some of the things we are still working on like interactive capabilities and so on. And then we have the engineering team. Engineering team I am prematurely working with is the language engineering team and the tools engineering team. So we have people for that obviously in Redmond and we have the tools team for example like Visual Studio and VS Code team in our Beijing office. We have some people that are working on some stuff in Serbia, more on the interactive side. The store side of this is in India so we are kind of wide geographically dispersed as well. And the interesting part of it is obviously that we are not just build the Azure Data Lake service we are also building and running and supporting the internal big data service called Cosmos inside Microsoft. And that means we are daily answering questions and helping with issues, etc. that our internal customers have in addition to build that external service. And that’s probably something that is kind of unique in that sense because let’s say if I was doing SQL Server I was having a box product. Or maybe an Azure SQL Database and I didn’t really have like an internal business running directly on the infrastructure that I was responsible for. I guess Microsoft IT would be using SQL Server but they have their own database administrators doing that. We are the people that run the internal big data platform for Microsoft as well. So that’s kind of where it is a big difference there.

KN:  The Microsoft employees are testers…

MR:  to some extents yes, as well, yes.

KN:  … are we talking about the team of 4 maybe 10 people?

MR:  So the PM team currently on the ADLA side, all up with everybody is probably about 30 people. 30-40 people. If you are looking at the engineering side, just the language side, I would assume it’s about a 100 people maybe depending whether you call counts of the tools team etc. So it’s definitely not as big as SQL Server.

KN:  But still is a big one.

MR:  yes, pretty big team. But you need that because like we have to keep our service running internally as well as do innovation on the external service and internal service as well, so you need a certain critical mass of engineers to be able to deal with that.

KN:  and also the product must be developed very fast.

MR:  Yes, right now we are trying to do a half yearly planning sprints and then basically try to shape about every 4 to 6 weeks refresh based on what we are doing on that. So yes, we have to be pretty quick. Now some of the features still take a long time like our partitioned output which is the frequently asked question on our feedback page after I think if you count all together the interactive parts. That feature was in development for a while but because it’s a fairly fundamental re-architecture of fundamental assumptions that we had earlier it’s going to take a while. And lot of components have to pull the work together. We are getting close to release it, hopefully early next year (this was recorded in November 2017), so yes, even if you have a 100 people at the end many of those are working on different parts. At the end you have one or two people working on one particular aspect so it’s still taking some whiles.

DW:  Yes, actually I was just thinking how big the investment is? Because I can imagine that 40 managers plus let’s say one team is around 100 people and there is more than one team and we see this from the perspective from one or two features on the portal. And we cannot see how much work it really took you to deliver this feature and this is really impressive.

MR:  I mean the point here is we obviously want to make it easy and making it easy means that we have, there is more cost to it for us because making it easy is actually harder than just putting something out and you have to do the final mile yourself.

DW:  That’s right. It’s easy for us to use it or for our customer to use it so you have to pay for it to be. You have to do a great development on the other side to make it easy. That’s great.

KN:  At the end of our conversation tell us where we can find you. Where the people can find you on your blog and information and knowledge…

MR:  so basically the best way to reach me obviously is either we have an email alias that I have created specifically to reach out for Azure Data Lake Analytics and USQL access which is just usql@microsoft.com

KN:  I told you Damian, I told you…

MR:  My Twitter handle is @MikeDoesBigData, quite simple, although it is a long twitter handle and otherwise I try to read the USQL tagged questions on the StackOverflow and the MSDN forum on the Azure Data Lake forum. So that’s kind of where I am currently hang out. Otherwise you might try to catch me at events like SQL PASS, Ignite and other places. But these are the kind of the main places. We do have a blog on MSDN, I have a personal blog there at MichaelRys is that. You can find me there. Or the Azure Data Lake blog is where we kind of hang out and post stuff as well. And if you have any like feature requests or questions or so feel free to reach me out so I try to answer all mails that we get.

KN:  Great.

MR:  Sometimes it takes a little bit trying to catch all emails.

KN:  Ok, thank you very much Michael this discussion.

DW:  Thank you.

MR:  Thanks for having me and I am hoped to see you either diving into the Lake or at least skiing on it using SQL Server.

KN:  Yes, for sure! Thank you very much.

MR:  Thanks, bye bye!

Useful links:

Michael’s Twitter: @MikeDoesBigData
Michael Rys on MSDN
Michael Rys on Channel9
MSDN Forum: Azure Data Lake