Metadata Management is Key to Data Governance Initiatives

hello everyone and welcome to today's presentation metadata management is key to data governance initiatives I would like to start by introducing you to today's speakers Irene polokov is the CEO and co-founder at top quadrant we also have Jack's the Box Jack is a semantic Solutions Architect at top quadrants in today's webinar Jeff will introduce the capabilities needed for effective metadata management it will outline how top grade edge enables a comprehensive but incremental approach to metadata management and overall collaborative data governance by providing a suite of modular packages that can be combined to support your targeted scope of data governance Irene will then demonstrate metadata management capabilities provided in top-grade edge she will also illustrate how edge integrates metadata management with other data governance practices such as management of business terminologies and data warehouse reporting hierarchies understanding data usage and data requirements and collaboration across all data state stakeholders before we get started today let's review a few logistical items please feel free to ask questions as we go along by using your GoToWebinar controls to post questions please look on the right hand side control panel you should see a drop-down box for questions if you click on the box you should be able to post your question at any time during the presentation we will address as many questions as time allows at the conclusion of the webinar we will also have a couple of polling questions throughout the today's webinar and we'll be recording today's webinar as well as send a link to send a link to the recorded version with that I will hand it over to Jack to get started all right well thank you very much Kristy and let's dive right into today's program on metadata management a few words first about top quadrant we believe that capturing the meaning and information is essential and our strong commitment to standard-based approaches to data semantics makes this possible our mission simply is to make enterprise information meaningful today's webinar describes one important way to do that by managing metadata and creating connections across silos we work with a variety of enterprise clients across different industries and several customers in the financial services sector aren't shown here financial services significant focus area for data governance and metadata management because of the complexity and its compliance requirements we're going to be using an example from that industry today so to get oriented before jumping into the product demo I want to talk briefly about metadata what is metadata why is it important and what are some of the challenges involved in managing it and especially how does semantic standards fit in to help address these challenges afterward Irene will demo how we do this in top right edge and how edge helps so what is metadata the simple answer is it's data about data and you can see some typical characteristics people capture about data here the origin definitions security policies etc but actually it's not quite so simple because sometimes one person's data might be another's metadata and vice-versa and more significantly it's related to context there's many different contexts you can see here examples in business technical and operational context and there's many different associated types of metadata what's important is that you can capture all of it and use it to drive value in the enterprise but before I get too deep into metadata I want to back up for just maybe one minute and say a word about meaning it's probably fair to say that all of us here are involved with information and decision-making in one way or another and it's probably fair to say that we could agree the good information in general is better than bad and of course we all know that's easy right well or not so much we've all made bad decisions and some of us have even learned that we want to make better ones there's a saying that knowing the good information that a tomato is it fruit doesn't really mean that we should be putting it in a fruit salad at least not if we want the intended outcome so there's something here it's not just good information but meaningful information that's needed if I hear thunder and see dark clouds inside this might mean a cup of tea in a good book outside means something else entirely in a classroom this might mean one thing here quite a different meaning and we can't make good decisions about what to do without information about the information we're receiving so meaning is about metadata making meaningful metadata available metadata that's accurate timely and relevant is critical for good decision making so I talked about why good information isn't enough for good decision making and that meaningful information is needed well why is metadata so important it's because the meaning that's so important is actually carried in the metadata metadata answers important questions like the ones you see here who uses it what are the limitations of its use etc and the meaning it provides cuts across departments systems and silos managing this metadata means more than accurately cataloging values for it it means making its meaning useful that is available timely and correct for stakeholders needing shared understanding of enterprises information so here's an example and in this example metadata describing which fields are personally identifiable information is missing and causes confusion this is a major area of work for enterprises a worldwide now because many of the regulations that are getting to the place like gdpr in Europe and the people in this illustration don't know what data to protect or the risks of non-compliance they're missing the answers to key questions on the previous slide they don't know what the customer ID or other fields mean in the context of regulatory compliance they need meaningful metadata to make good decisions and a shared understanding of the meaning is essential to its effective use and governance so how do we do that that's where semantics comes in and a top quadrant we're deeply committed to semantic information management semantics is about capturing meaning to make it available for good decision-making simply and flexibly just like ordinary sentences semantics describes the things that we care about and how they relate to each other things and relationships its language having a common language makes communication possible and semantic standards are a common language semantics standards were designed to represent two key types of metadata and if you're going to have a metadata driven solution to make your data meaningful then you better be able to capture both types we need to understand what something is and what something relates to well if you want meaningful metadata the semantic standards were designed just for that today as a result of a 15 year effort by the w3c a collection of standards enable the capture and communication of meaning how well these w3c standards stack together in a system that allows us to say things to share a common vocabulary to infer some things about what we're saying to be able to ask about what's been said and finally to be able to verify all of this with rules so given that this is possible oh I managed your metadata semantically well here are just a few of the priorities that require correct and meaningful metadata for these purposes that you see here we need metadata to connect the business technical and operational aspects of informational assets into a single data landscape but there's many ways to do this and many systems for doing this and some are not as complete as other semantic metadata management focuses on the relationships that assemble and glue together today's enterprise information landscape the standardization of semantics finally allows us to capture meaningful metadata and perhaps more important to talk in a common language about our data across systems for business understanding well if human beings are the glue that connects this variety in the data it's not going to be easy and beyond that it'll be expensive slow and error-prone metadata can vary with different information departments and business needs or more and some metadata may not even be known yet organizations must have systems flexible and extensible systems in place to capture meaning of all types of data today's data and tomorrow's data and this is where top right edge comes in edge provides flexibility and meaning with semantic metadata management how do we provide this capability and edge top right edge uses the simple notion of assets to capture and organize diverse Enterprise metadata this pre-build and easy to understand framework provides standard asset modules for business technical and operational metadata edge then allows you to map your organization's actual diverse information assets to this rich standardized data model these mappings organize your information for understanding and they can be easily extended or customized as needed with modular packages that you see here on the left so in just a moment you'll see Irene use edge to do exactly these sorts of mappings to make meaning available in an actual banking example as a holding company top bank corporation needs to report quarterly to the Federal Reserve about its financial health the Federal Reserve's Form FR y9c and it's supporting schedules just like our tax returns collect information about income and assets top-ranked Corp has various databases with tables and columns containing data about its holdings and Irene will be showing how edge captures descriptive metadata in standardized data technical and enterprise asset collections more importantly she'll be showing how edge associates metadata relating data elements to the forms and software they use them for top Lankes compliance activities by managing relational metadata the meaning of each step in top banks reporting processes preserved and top ranked can ensure accuracy and the lineage of their reporting but before I hand you off to Irene for the demo we're going to take just a few seconds here for a quick poll Kristy thank you jack as I mentioned earlier we would appreciate your feedback and a few quick polling questions we're going to do one polling question now and we'll do two others right after Irene's demo we will share the results of each question so that you can see how your peers responded so if you'll please take a look to your screen to answer the following question thank you jack what does your organization use today to collect or manage metadata please select all that apply spreadsheets modeling tools for business processes er objects for example ei Arwen Embarcadero catalogs within an ETL such as Talland data cleansing tools such as Trillium or bi products such as Spotfire metadata repositories such as Rashaad or adaptive and finally integrated data governance solutions so we'll get a second for everyone to respond okay it looks check it looks like everyone they have entered a response so if we want to close the poll okay Christy and share the result there you go interesting um 78% our spreadsheets and it looks like modeling tools and catalogs or the second and third option thank you so much for taking that foot pole and we will continue with Irene's demo Irene Thank You Christy dollars presenting do you see my screen now yep all right let me start by re-entering cue a little bit in this environment what you see is a homepage of tab right edge as Jeff was describing it's a pretty reach and comprehensible comprehensive free require environment and we managing it different types of assets which are organized in collections of like asset types as you see on the left so you know data asset collections for example have information about databases data sets etc groceries organized glossary terms Enterprise Asset collections that you see things like form schedules made the organizational capabilities and so on technical assets collections are about applications infrastructure servers all of this different information connect together because the meaning as we talked about it is in about connections and table right edge is a highly configurable environment so this types of collections that I that we see here can be configured to show more or show less in fact I'm not showing everything and new collection types can be added the other things that I'm seeing here are tasks that are assigned to me and then task gets created to support collaboration and workflows I get notification in the email but I also conveniently see it here on the home page let me go in data asset collections and see what kind of collections I have here so our example will be focusing on tablet Bancorp so I see that collection but I also see some others I am logged in as administrator and as administrator I could manage everything but in general this is organized by the kind of things that you could see the kind of things that you could edit and the kind of things that you could manage in fact in a different browser I am logged in as a different user and you see that that user I only have access to two data collections so this is a mitigation of view that is based on the type of collections I could also filter it by tags that indicate that it belongs to a certain data domain or whatever other categorization we may want to have and if I wanted to have a cross-cutting for you I could click here on a fine asset collection in fact I have it open in another window and here you see that I have a different collections that field is banking and some of it is about enterprise assets such as forms and schedules you know groceries my data assets etc if I wanted to have that kind of you as my mainly in the system I could do that as well and here we also see some metadata this is a metadata that is actually about collections themselves such as name identifier who is responsible etc so it's a different a different kind of metadata but let me close out of here and go back to go back to to my data asset collection and take a look what title data assets I I have there so as I've already mentioned this is a data databases data set even logical models like logical attributes and and entity and of course physical models such as tables columns etc I have a few databases here and if I wanted to look at all the columns across everything I could do that which which is what I just did in the navigation so you see there is quite a number of them and then let me look at one of them it's an account number so far the kind of metadata I see here is ready I keen to the data dictionary and then you know but it's a primary key physical data type and so on and that's natural because it's actually was imported from a database and that type of information they have I could navigate across going to the table that this column belongs to and see it has two columns and if I go to the customer name we see some additional metadata the metadata that is managed only by this environment of course it cannot come from from the database itself things like to technology data quality mapping geological attributes so here we see one type of connection and there's even more metadata field available to me I could see them once I click on edit but actually I have a tasks the task that's assigned to me to make some changes so why don't I go back to my home page and actually work on that task which is asking me to connect this securitisation exposure field to a glossary term so this is where I am going to click on edit and actually make my edit and also see what other metadata is available to me it's quite a teach collection of things that have to do with statuses with documentation compliance traceability you know governance in terms of who is responsible accountable etcetera for for this particular data elements with it that it's critical and so on so what I wanted to do is map it to the glossary term and if I don't know whether the glossary term where the appropriate glossary term exists or not I could of course go directly to the glossary and check it there but I could also search for it and let me see so what we have above securitisation and in fact we have a few things and here's one that seems to be the right one so I'm going to connect to that or I could have used an autocomplete and just started typing AFF securitisation and the pulldown list would come up for me so let me save that change and you see that it maps to the glossary term also if I wanted to by the way all my changes are or the trail so if I wanted to see a history of my changes I could I could see that and see that this was just added by me on on a date and I could alert the change and and so on so we saw that there is at each amount of metadata available and all of it is included in the models that are shipped the stub rate edge the models are very reaches you've seen and however when you use it in any organization no matter how much you think through those models there will always be something special an organization may want to capture that we have not thought about something that's unique to them something that's different maybe some of the fields that are available here notably organization wanted to use that so H is a highly configurable environment and the way it's configured is through ontology models that's another type of asset collections see oncology's classes properties and I could navigate here to the anthologies any intelligence that I manage including anthologies that actually configure and customize metadata integrate edge itself so I have I have one here already but I also have it open in the window and I'm going to go there so this is metadata about my metadata so essentially a meta model of top right edge users WCC standards that Jack has mentioned and here are all the different types of data assets I'm going to make changes to the columns so I could I could look at up that class of columns so here is um here are columns database columns and some information about available about it I want to add a new field so there is already some identifiers but let's say my organization has some specialized identifier that is additional and I'm going to create a field that is just called special identifiers Eyre and let's make a committee chair to save that now what I want to do I want to place it in some section on the form and the way I do this is by creating this shape for it for that and in the shape I could create some additional information so for example I could put it in the property group those are the sections on on the form and I could one say restrict its value for example say that it has to be between ten thousand I'm going to say that change and go back to where I was and see if I could use this new field I'm going to click on edit again and say did I do this correctly [Music] interesting I tried it three times before it it must be a damaged engine but let me let me assure you that it doesn't it work and I should have had a new column appear so I'm going to investigate why I didn't buy the initial but in the meantime let me see what other connections I have so this this particular field a particular database column is used in in the technical asset collections and I could I could see that that it's actually used as a relationship in in a program which is in the technical asset that an SQL script so I want to see more about it and I'm going to click here and navigate to my technical asset collection where I just have three different programs so you see that this program has inputs and outputs and it's an SQL kind of stored procedure and I could have information about it here and I could also explore since it's about inputs and outputs I could also explore it visually in this neighbor gum type of visualization it shows it in the diagrammatic view so we see that this SQL scripts have inputs it has some outputs and if I wanted to clean up this view a little bit and don't show the type or executable type I could do this and I could further explore relationships and just walk this graph and kind of connections in through the visual navigation so that's another kind of information that is captured here about applications and how they use data sources as inputs and and output yet another type of information is enterprise assets such as forms business capabilities and so on I I do have one of those open here so let let us to take a look at it this have various schedules that our fictitious bank uses to report information to the federal Federal Reserve Board so we see this is one of the reporting forms FR y9c it has a number of schedules and I could navigate to those and see one of them and also see if it's used anyplace because one of the important questions for me could be oh well how is that schedule actually produced and it pretty the way it's produced is captured in a lineage model that describes how various inputs and outputs flow through programs and applications and produce that schedule so I have that connection available to me as well now one of the things to mention about this environment that it's it's primarily for the data stewards or people involved in managing metadata technical architects business architects and and so on there's search capabilities across the environment and this could provide a simplified view to publish information to the broader enter so if I wanted to just search here for let's say securitisation I could I could get a search view and that view is something that everyone's not just data stewards but everyone in Argos organization could use we could configure what's available in the search and this allows a simplified view across all the assets that are managed in this in this environment so what I could do is click on it to see more information what kind of information I could see and in general how search works is configurable here as well so let me try that configuration and hope that it works better for me this time so if I wanted to configure configure my search I could set what collections are available for searches I could set how they available to search I could say set what shows up in the detail view here let me just add additional fashion for example physical data type is a fashion and if I go back to search and just refresh it see what happens so you can see that now I have another facet available and there are other additional configurations that I could make now this environment could be used not just to see information but also make comments about information and if I have appropriate level of access I could drop down to to the type of information that to do really much more detailed view like for example if I wanted to see this glossary term within the entire within the entire glossary I could navigate to it from from there all right there we go so here I went to to my glossary with my glossary terms I could see them here are key claim in the three different types of exposure including securitisation exposure so there is a variety of the access points and navigation mechanisms that are available here and as I mentioned before from the search environment we could also make comments so if I'm going to make a comment here but I could I could have made it from the search environment so if a broader enterprise users who only have access to search wanted to say something about let's say definition of of this term is incorrect please take a look at something they could do they could do that here as well and that comment goes to goes to the stewards and they could then see the comment here and make appropriate changes initiate the change workflow and complete their job is needed there is also rich metadata about the attributes about the glossary terms themselves so if I wanted to click Edit here we could see what kind of metadata is available for a glossary chairman you see there's quite quite a bit here awesome Jack was there something else that you wanted me to show because I could I think I have a little bit more time so if there is anything else that I need to show I could quickly do that to show that the glossary were in general for the demo more general for the demo oh I mean what about importing sorry jack importing are talking about importing someone without okay very yeah very good question Christian and yes so let's take a look at how I could import data there is a variety of important are they available to me and they depend on the type of the assets for example for the data asset collections I could bring data from the DDL file from JDBC collections from from a stretch sheet for other types of collections you know importing it from VDL wouldn't make sense so there is variety of configurable importers there is also a variety of exporters that I could use in in for data asset collections for example I could export JSON schema in and and use that to to produce avril for example for the big data asset could you talk about glossary specifically Irene the primary imports would be from from your spreadsheet and you could also import RDF files into glossaries as well all you could import the spreadsheet and there if I had a spreadsheet to be selected I could then map the columns from the spreadsheet into properties for my glossary model so the import capability is actually quite quite configurable so if you have any spreadsheet irrespective of its structure we could always map the columns and import not just relationships not just not just attributes but also relationships that are being used so with that I think we'll probably pass back to Jack to maybe show some of the additional features that I didn't cover and also to answer questions ok thanks very much Irene and before we do that just going to run another couple of quick polling questions here and Christy thank you jack please take a quick look to your screen and we'll just do two quick polling questions before we go back to Jack's to finish out a few slides so the next polling question you should see on your screen what best describes how happy you are with your current state of metadata management so you'll see several choices unsatisfied with the current solution but you don't have any plans to improve it unsatisfied but expect to start an initiative to improve it in the next six months unsatisfied with the current solution and evaluating alternatives right now currently implementing a new solution are satisfied with your current solution so if you just want to okay I think we're good to wrap that up okay if you want to share as a result Jack again okay so a little bit all over the place several unsatisfied are currently implementing and 28% are satisfied okay let's go to the next polling question briefly Jack if you don't mind sharing them okay New York okay what is your biggest issue with your current metadata management capability so which of these are you not able to do well currently so see an integrated view across metadata for all sources including structured and unstructured data bases and big data lakes connect all types of metadata including technical business and operational easily extend the environment by creating user defined attributes new metadata types and api's ensure access and collaboration across technical and business stakeholders or use metadata for impact analysis to facilitate integration or to identify redundant sources etc I just give a couple more seconds folks are still I mean their response [Applause] okay all right Jack I think you can share Chuck could you pass back to me actually I figured out what was wrong with my customization so I could show attributes working and sure has the value range working as well okay so I think everyone's had a chance to look at the results of that hole so thank everyone for taking a few seconds to do that for us and I think we'll hand it back to Irene thanks Christian yeah yeah so Tim what was the problem was that I actually didn't include my customization so which I did now while we were doing those polls so you see now that the special identifiers showed up and let me try to use this field I am intentionally going to make a mistake and enter five here because remember I had that rule that it has to be between I think ten and a thousand so let me try to save the change and you could see how that rule is working so there is a data quality and protection aspect to the environment as well it will not it will try to prevent me from entering incorrect values so I'm going to cancel out of here this add jack back to you okay um thanks very much Irene I'm glad you were able to get that taken care of coming in talked a little bit about metadata in general Irene's showed you edges ability to handle metadata edges metadata is actually situated in a larger governance framework as a result there's a lot of capability in edge that we don't even begin to have the time to cover so I really painted out that she was administrator and showed what it was like as a different user in a different browser interface just pointing out that because it's a web-based system of we naturally have role based access controls and this goes down to a very granular level to the system as a whole or to the individual asset collections that we've been working on we also have both production and working copies production is what you might expect if the metadata that you've been managing needs to be made available other people or systems you want that live and running on a regular basis but if you need to be editing it or developing it or putting it through an approval process we have these working copies which are virtual copies of the production vocabulary created specifically for editing purposes and those working copies can go through an approval process can be frozen directed proved and committed to production and this can be cyclical it's very very flexible on the entire platform not just the metadata management offers full multilingual support and we have dictionaries running in the background you can see here in the example that we can have preferred labels in a variety of different languages Irene just showed you a bit about our data quality rules and validation and without getting into it too deeply these apply to all of the different kinds of information governed by edge not just metadata and these can be their standard ones like this that came with the product or they can be configured entirely from a custom perspective with business rules that could apply to different departmental policies and security policies and all the compliance of your data with these policies that you might create can be reflected in dashboards inside the edge software we look at completeness how the descriptions changing the validity over time see where you are in your processes and then finally and this is a huge huge area edge allows for because it's a server based product it allows for integration with other systems so it may not just be individuals who are looking for metadata about information um there may be systems that are requiring that you can see on the lower right here you could have this integrated with a data warehouse with portals like search portals other in-house applications for example in finance or sales and all of the data is centrally stored in edges repository the screen that you see really is just a user interface interacting back and forth with the repository that is part of the server system and as a result we can access that directly from third-party systems through Sparkle query access or through Java messages or directly through Web Services many of which are automatic in the system so enough about that there's an enormous some out here available in the details if you want to hear more about that just to wrap it up top rated edge uses assets these data assets and technical assets to capture and organize and enterprises diverse metadata in a pre-built easy to understand framework stores a variety of this descriptive metadata and even more important it captures the relational metadata for assembly into an entire landscape we do that with packages we mentioned them casually before each of these contains different assets that you might want to work with and they support an incremental approach to governance you can start with any of them and add them in any combination in summary top right edge because of its semantic foundations offers an enormous amount of flexibility and extensibility you saw Irene add a piece of metadata that wasn't even thought about initially these models we have both pre-built and user-defined models and we can make connections between all of these so with that I think we're about ready for questions Kristin thank you jack and if you haven't already please submit your questions and discussion points in the question box located in the right hand side control panel we've got a couple of questions here that we're going to start with first one here is it seems like you have a lot of different types of assets and some different packages to consider where are with what do your customers mainly start TD Irene well this really differs and we have some customers who have very precise goals for example they want to manage reference data and they just use that package and later on they could expand or they just want to do business glossary so they just want to do metadata and that's the package they start with they could always add to that they could create also their own asset collection tight some customers are not quite sure where they going to start so they just license a combination of packages and then they could enable and disable and combine them in any way they they'd like to so there is a lot of flexibility and this lot of variety in the customer use cases depending on what exactly their focuses I can say that there is one determinate pass and that's fine because the system is flexible enough to support whatever journey one starts with all right Thank You Irene let's take this question you showed creating data assets within top braid edge what happens when database schemas change that's a very good question so I showed some import capabilities that Christy asked me about and you saw that information gets imported from the database schemas themselves typically although some customers may want to import this information from the modeling tools such as urban if they if they do logical and physical model and those tools but certain combination of places that they import data from now import can be done on repeated basis because metadata does change I mean database schemas don't change quickly but they do change so when you run the second import what gets imported is Delta so if something has changed let's say a new column has been added then that information gets imported now everything that was entered on a by H itself such as this additional metadata that I was showing like criticality personal and information indicator or how it's used in applications all of that gets preserved so the the subsequent import do not override it in any way they just only bring with them the changes that came from the source environment Thank You Irene and here's one other questions that I think we should take someone's asking how we how we are different than Calibra and I think that's a great question it's probably something that requires a more extensive response and we'd love to continue to chat with you about that however one key way that we are different is that we are based on semantic standards Irene I don't know if you want to talk a little bit more about you know what that means I know we've covered some of that in today's presentation but I think that is a key message that you know we want to get across is what semantic standards can do for you and that is that is flexibility and customization Irene would you add anything else to that well yes so it is based on standards and there's increasing claim organizations in all industries are starting to use w3c standards for developing their own custom anthologies or you're using industry standard technologies for example fibre in the financial industry and because it's standards-based you could just very seamlessly and add it to our environment and sosa means that variety of other standard interfaces are available in terms of de-referencing dereferencing any item any asset that is managed in integrate edge by that we mean using its unique identifier its uri to fetch information about it and to go from the information about one asset to the information about another asset also you saw me blow it first unsuccessfully to extend the metamodel behind behind this environment and in doing that I was just using standard ontology modeling approaches where I create classes and I create properties and I connect them so all of that is standards-based you don't need to learn a proprietary in an object model object modeling approach that may be in the tool such as calibra and interoperate is anything else that is standards-based and you know you could use you could reuse skills so if you have ontology model modelers who are trained on these technology you could use that and the other point of it that the environment has been built from the ground up to accurate in a heterogeneous setting it's very flexible it has at each set of extensible models behind it that covers a very broad spectrum of assets asset collections and reach even in a sense that you saw me for example adding crew so that about the identifier being between a hundred and a thousand that rule language is also using w3c standards there is a creative language where all the information can be created and you could create template those queries to create web services that's also based on the standards create a language called sparkle for creating graph databases so importance of standards how we leverage standards flexibility richness when you use standards like that there is much greater chance that lots of things have been thought through for example the type of rules that I could add the overall scope of the language is is quite quite broad because it's based on the requirements from many organizations of the kind of metadata and information they want to capture and connect okay thank you Irene we are pretty much at the end of time so I think we'll go ahead and wrap up today's webinar and we had quite a few questions and we will be responding to those questions to look for that in your email for that response and it also has mentioned early on their presentation we will be providing a link to the recorded presentation as well as slides thank you so much for your attention and your interest in today's webinar and we hope you have a wonderful day thank you [Music]

Channel: TopQuadrant

Views: 2,376

Rating: 5 out of 5

Keywords: data governance, metadata, metadata management, webinar, topbraid, TopBraid EDG, TopBraid Enterprise Data Governance, governance, data management, semantic information management, semantic standards, semantic, semantic data governance, compliance, regulatory compliance, business terminologies, governance practices, data definitions, data assets, technical metadata, business metadata, operational metadata

Id: gOC6KyZgfmw

Length: 56min 7sec (3367 seconds)

Published: Tue May 02 2017