Effortless database management with AI

Key Takeaways

[Music] I'm n bagat i head product management for Gemini databases with me on stage today you'll have Alexi oin product manager ginine databases Katherine Colman ux director and frontend director for databases Bogdan capatin a technical expert at Ford it's 8:00 a.m. your first session in Las Vegas I know this audience is truly interested and excited about how AI will transform database management you know we're going to make this session engaging and exciting how many live demos do you want to see today with Gemini one two let's do five live demos five live demos with Gemini but we let's have feedback loop you know every time a demo is done I'm hoping it all works out with Gemini I need to hear a round of applause right and if you love it louder the Applause right are we ready all right we are three-part agenda first I'm excited to introduce Gemini in databases what we announced yesterday second we'll doing we'll be doing product overview and demos and third we have Spotlight session with our design partner Ford as you see we have a pack agenda so let's get started every organization we talk to tells us managing databases is getting more and more challenging we believe there are four key trends in the last few years which is causing this the first one is this proliferation of databases making it them hard to manage second the responsibility of database observability and performance is lying on the shoulder of devops and Developers and they are struggling to keep up third as part of modernization Journey more and more customers want to move from the legacy on Prem databases to Cloud managed databases and those migration Journeys are not easy at all and fourth the rising security threats to databases the strengths of made database management complex once you zoom out and look at the Modern database management as you see in the picture you start understanding why the three reasons you can observe on this picture number one there's unprecedented scale database instances per developer per admin have been increasing at rapid Pace in last few years second a modern database Fleet does not have just one database engine type but have many many database engine types and each of them can be different from others and third there are number of personas who are involved in managing this database Fleet uh you have generalist personas who are not database experts and then you have database experts who are specialist and then you have horizontal personas who are responsible for inter Fleet and some responsible for just few applications the tools available for all of this personas today tends to be Point Solutions fragmented and often hard to use no wonder database management is getting challenging by the day how did we get here to understand how we got here we need to look at the evolution of database management for last 15 years it has evolved from self-managed databases to manage databases such as Cloud SQL memory store you know it it it was a magical experience for the first time when a developer sitting on a chair could press a button and get a MySQL instance and eliminating ton of management things which they used to do this was the first pivotal point when it came to data management then as the demand for Rich and more reliable applications grew we had Google introduced and Unleashed Cloud optimized architecture such as Cloud spanner Cloud big table alloy DB this is a second inflection point of database management it reduced complexities such as sharding databases horizontal scaling multi- region availability each of this Innovations have been a step function when it comes to database management reducing unit cost of building and managing databases dramatically as a result you have more apps more richer apps more frequent releases and you see the complexity which you saw on the previous slide we believe we are ready for the next pivotal point in databased management the one powered by AI to tame the complexity of volume variety and velocity of modern database management needs I am so excited today to announce the next evolution of database management with Gemini in databases it is your AI power database assistant which will transform how you build manage and operate your databases it will dramatically reduce the unit cost of building and man manage databases allow you to build more and quickly starting today we have announced five new capabilities when it comes to Gemini and databases first enables you to streamline and expedi database management second improves your developer productivity when it comes to database third with d risk and help you optimize your database fleets even of hundreds and thousands fourth will enable your developers to stay ahead of performance problems fifth will enable you to have proactive governance all five of them are available for all of you right now to use and try so let's look at each one of them in details starting with databased migrations many of you have been involved in database migrations and you know they're not easy not easy at all a seemingly simple migration can take up to 6 months and complex migrations can even take 2 years in our experience there are two things which comes and adds challenges to this migrations first is code conversion especially for heterogeneous migrations where a source and destination have different dialects and code conversion and production stages are at high of high risk and error prawn migrations are time consuming as they required repeated task over your hundreds and thousands of lines of database code and again you have to make sure you got each one of them right to address this we have introduced Gemini in databases capability which is Gemini assisted code and schema conversion in our database migration service with Gemini powered code and schema conversion it will automatically help you convert your simp simplify your last mile of code conversion a cross equal dialects for example from Oracle to alloy DB in uh in this case we know migrations are tricky so our AI just does not even help you convert but it provides the explainability so that your developers can understand why certain conversion is happening and suggested and making them familiar with the destination dialect such as pogress SQL i' like to call upon cat to show Gemini databases in [Applause] action all right thanks NES so I'm going to walk through a fictitious example of symbol shops to illustrate how Gemini in databases can simplify your entire workflow so symbol shops is a 50-year-old big box chain store and and they've been interested in migrating the remainder of their on-prem databases to the cloud where their newer applications in infrastructure already sit their CIO has asked Dena their DBA to migrate their legacy databases from Oracle to alloy DB starting with the electronics database in order to complete this task however Dena needs to understand the source and Target dialects well and she doesn't let me show you how Dena can learn to fix conversion issues using Gemini explainability in database migration service let's go to the demo okay now Dena is in her Oracle to postgress uh conversion workspace but she hasn't done anything manual yet so there's been no conflict resolution at all and and she's looking to resolve some of these conflicts she clicks on one of her stored procedures and she has a handy conversion assistant right here in the top right where she can ask how should I fix object conversion issues and this is really cool because Gemini understands what the issues are provides her with a refactoring approach an explanation of changes some key points and probably the coolest piece is this side by side view of the conversion so on the left is the code after we did an initial Oracle to postgress conversion and on the right Dena can see what Gemini has done to refactor the code this way she can see exactly what Gemini did she can review the postgress and most importantly she can learn from it so Dena is going to go ahead and insert the suggestion save the squl and she's all set to go now once she's done this a couple of times Gemini will take over and will apply these changes across all of the conflicts in her conversion workspace now that simbol has migrated this Legacy database the next thing at the top of their minds is how to accelerate application development back to unes thank you CAD that's the feedback Lo thank you kind of cool live demo developers are one of the most critical assets to every organization making them productive is the highest leverage task any organization can achieve with databases developer productivity is not easy a recent 2023 stack Overflow survey said 82% of developers spend more than 30 minutes each day searching for Solutions with databases we see this number even to be higher in our understanding this happens because a lot of developers today building databases are not database experts they full stack developers application developers they do lack sequal expertise and the same time today they are asked to they tasked to code SQL in multiple SQL dialects making this job even harder the tools available to them are complex today every organization wants to empower the tens hundreds and thousands of developers but they want to do this in a secure and compliant way to address this I'm excited to introduce database Studios with code assistance database studio is an in context code editor available right inside Google Cloud console it requires no setup with database Studios developers have access they can test and run SQL only when they have the right permissions now with Gemini a user can simply code SQL using natural language they can further High let a complex equl statement and get their get this understanding in natural language such as English let us look at how Simba shops made their application team Innovative with Gemini over to you cat all right thanks again NES so it's almost the end of the quarter and symbol shops wants to run a flash sale on their excess inventory pretty common use case right so they've asked Anita their application developer to ensure the flash sale tag seamlessly displays on their app to attract Shoppers let's check out a demo of how Anita can do this using alloy DB studio all right so she has navigated to alloy DB studio in our Google Cloud console and she's gone ahead and gone to the electronics database and let's click on our editor tab so we can go ahead and open this editor Anita is still getting used to writing SQL and so she clicks on this handy Gemini code assistance for help and she's going to do this prompt here list inventory where stock exceeds 50% average daily sales oops let's spell sales right aha so she goes ahead and uh sees that Gemini has inferenced the electronics table and suggested SQL code for all excess inventory so she clicks insert to insert the query she wants a little bit of help understanding the SQL that Gemini has generated so going to go ahead and select this query here and we're going to click on our little Sparkle icon which goes ahead and opens up our Gemini chat interface and so the Gemini chat interface is now going to break the query down into easily understandable components so that Anita has a better understanding of the database structure and the tables the query is pulling the data from because of this Anita feels pretty confident in testing her query and verifying the results and she is now ready to go ahead and update the app and so she runs the selected query now that the application team has been able to update their app the next thing on symbols wish list is to see if Gemini can assess and prevent risks before the sale goes up back to you NES thank you CAD wasn't that cool when organizations are moving fast and they have hundreds and thousands of databases the next top of the mind question for the database platform teams is how can they stay ahead of risk however with modern database fleets this is not easy here is why a Redgate survey in 2023 said majority 79% of it teams are now using more than one database platform and as many as 29% are using more than five each of these database engines have different terminologies different apis different ways they work making database management to be very very complex today further the lack of centralized visibility available to the customers makes en leads them to surprises and unexpected risk further it inhibits collaboration between all the personas we saw earlier this is again where Gemini can help we're excited to introduce a brand new capability database center today with Gemini in databases it provides a single pin of glass for monitoring the health of your databases across your organizations is across your projects across your engines it is AC cross database use cases without ever having to switch between tabs or login different environments now you have unified database management at your fingertips further Gina has proactive recommenders which constantly work 247 to to detect upcoming problems and suggest fixes the coolest part comes next Gemini in databases democratizes database management by enabling anyone to ask questions about the database Fleet in natural language imagine someone asking the question tell me all production instances in Northeast which have failed backups in last 24 hours and get a response in seconds something which could have taken hours or days writing complex scripts or waiting for a centralized team let us look at this in action over to you cat awesome thanks namash all right so back to our flash sale all of our teams at symbol shops are mobilized to prepare for it and Pat at the database platform team has a checklist of things that he wants to ensure so are his instances right sized for the heavy load are there security issues are the databases being backed up regularly are they set up for failover let's go to a demo so we can see how Pat can check everything on his checklist all right so Pat has gone ahead and navigated to our Google Cloud console dashboard and he's going to go ahead and open up our handy Gemini chat and he's going to ask are all of my production databases highly available Gemini looks up all of the databases in his organization filters them with the label produ ction and returns a total number that have no failover enabled he sees that he can go ahead and click over to database Center so he's going to do that and once he gets there he can see that database Center not only shows Resources with no automatic failover but it provides a 360 degree view of the entire database Fleet covering checks for availability data Protection security and compliance and over 25 recommenders from Gemini to see the affected resources Pat goes ahead and clicks on resources not failover protected and he is immediately zeroed in to all of these individual resources he selects the first one which opens our handy issue details panel this panel shows all the necessary information to resolve this issue in one place and what he realizes is that this resource is not set up for automatic failover because it's config configured to a single zone instead of a multizone so he can go ahead click here edit configuration and he's taken directly to the edit instance page for this particular cloudsql instance where he can select multizone and go ahead and save it pat can follow these simple steps steps to fix all of the other issues he's found with all of this information in one place he can easily run through his checklist which previously would have required him to collect information from many different sources Gemini has made building and managing apps even easier as you can see but can it also help us tune and detect performance problems that we always tend to run into backed in a mesh to find out thank you Cat like simple shops no one likes slow running databases especially during the most CR critical time for your businesses however keeping up with database performance uh troubleshooting and optimization is getting increasingly harder here is why application teams are shipping code faster than than ever before the Redgate survey shows nearly a third of development teams are now deploying database changes to production in a day or less up 50% from two years ago the next challenge is the person who gets on call when the slow running database is is no longer necessarily a database expert such as DBA so full stack developers devops engineers and they lack expertise I always say the Legacy database Performance Tools were built by experts for experts again this is the fourth place where Gemini is going to add this magic excited to announce enhan insights with Gemini Gemini continuously monitors your workload database configurations and system resources it provides you system tuning Rec commanders such as setting the right database Flags or not running out of resources when you're moving so fast query tuning recommenders look at your workload patterns over period of time and suggest you provides you suggestions to improve your query performance such as suggestions for missing indexes Gemini chart integration right in context next to you database insights provides assistive troubleshooting it provides in context explanation in trouble shooting and also guides you to potential root causes let us look at how simple shops save the day with Gemini databases back to you cat awesome thanks Nash all right so the symol devops team runs a test of their retail applications to ensure quality performance on the day of their flash sale they notice slow behavior and an inability to add items to their checkout cards so Dave from their devops team received leaves a page to look into the issue let's go to a demo of how he can do this using Gemini so first Dave goes ahead and navigates to the query insights page in his Google Cloud console and he can see very clearly from the database load chart that there is a spike in total execution time particularly in CPU wait time Dave isn't a DBA and he's not an expert in this so he wants to figure out how can I better mitigate CP you wait time so he goes ahead and opens Gemini and he asks how can I reduce CPU wait time in my queries Gemini references Cloud documentation trouble shooting guides and a variety of different articles that allows it to suggest some help helpful tips for how Dave can do this everything from using indexes to using a query cache Etc much of these tips are obviously about optimizing the query so Dave goes ahead and takes a look at his graph again Scrolls down a little bit and um is now looking at the list of top queries by total execution time he notices that there is a create index's recommendation on this first query here and that's actually perfect because that's exactly what he needs so he clicks on the recommendation and he is able to see it explain which indexes and tables need to be updated and also which queries will be impacted so he goes ahead and updates um with these new indexes having made these changes the team continues additional testing and with no new issues they are able to successfully run their Flash sale as we've just run through Gemini has helped symbol with migration with development with Fleet Management with performance troubleshooting so now their leadership team wants it to help with one more thing that really keeps them up at night can Gemini help them quickly assess whether their entire org is secured compliant and they are not running over budget let's see if it can do that back to you NES there's always one more thing on the road map right right we talked to a number of leaders in 2024 two things keep them up at night first is protecting and securing their most important asset data and second is not overrunning budgets in 2024 you know we increasingly see cyber attacks increase as reported by many industry reports such as checkpoint survey over here as AI gets more advanced it is also helping Bad actors database security configurations across so many databased types of engines can be extremely hard to get it right missing compliance and cost overruns are common problems because of lack of expertise and time and again the tools available today for each of these databases are Point Solutions and lack of centralized visibility makes it hard to to detect the risk and stay ahead of them database Center also provides proactive governance with database Center proactive recommendation scan through your risky configurations missing missing best practice and finding wasted resources in uh integration with Google Cloud security Command Center forther detects missing compliance and security most common security problems users can further ask Gemini quickly in natural language to understand the problems and understand a complex say compliance standard let us look at how symol used it used Gemini to protect his entire organization over to ukat all right so symol shop successful flash sale fueled rapid growth leading to new contracts and a broader customer base which is awesome but now they face the challenge of meeting a new compliance standard across their entire organization Ian is tasked with implementing the necessary changes to achieve this compliance let's go to a demo so we can look at how database Center helps him do this easily and seamlessly so first Ian's going to go ahead and move from symbol shops to symbol core and this allows him to get a view of his entire datab based Fleet across the whole symbol organization not just for the symbol shops project currently he can see by looking at the industry compliance card that uh nist 853 is not part of their regular compliance checks and he needs to make sure it is so he goes ahead he clicks on our handy customize button and as you can see he can customize multiple Fleet issues he looks at industry compliance finds the nist 853 and applies it now he can go ahead and click on the violations and get zeroed in directly to all of the resources uh that violate nist 853 he selects the First Resource and now he can see all of the reasons why this particular resource might be violating this standard what he sees here is that in this particular case this resource is violating the standard because it has public IP enabled he'd like to find out okay how do I disable public IP so he goes ahead and asks our handy Gemini chat here how do I disable public IP Ian has chat built into his workflow so he doesn't have to spend time digging through docks and looking for the answer to this particular question he can stay focused on his work right here in database Center and as you see uh Gemini has returned a stepped list of all the things that he needs to do in order to disable public IP Ian follows these instructions and fixes all resources that have this problem but while these instructions help fix the immediate problem he needs a longer term solution so he's going to go ahead and ask Gemini is there an org policy I can set for this he can seamlessly continue his conversation with Gemini getting answers to follow-up questions without needing to provide any additional context and so as we can see Gemini has now told him that yes in fact there is an organizational policy that you can set for this and Ian is thrilled because all of this uh happened extremely quickly before he would have relied on multiple scripts each tailored to a specific database or API and it would have taken him a long time Gemini and the database Center's unified compliance view streamlines the entire process now Ian is able to resolve the nist 853 issue in minutes so six months after symol shops finished its final database migration there happily running their full inventory in the cloud from migration to application development to Performance insights to Fleet Management that keeps them in compliance with a increasing array of Partners symbol feels com confident and secure in the assistance that Gemini provides to their organization back to you NES fly thank you CAD for wonderful demos five successful live demos with Gemini the demo gods have been gracious on us today you have seen how a fictional company symbol uses Gemini to transform its entire database management Journey let us look at this in real life we have with us Ford our design partner who is using Gemini has been testing Gemini to manage its thousands and thousands of fleets of databases I'd like to invite Bogdan capatina from Ford and Alexi oin product manager gini [Applause] databases thank you very much tames that was an outstanding five demos right next slide please sorry about that all right so while developing Gemini in databases we have been very fortunate to have Ford as our leading design partner in the process and bdon capatina is here today to share how Gemini and databases will impact Ford's database efforts so that Ford can maintain the that amazing level of database Innovation can you tell us a little bit about your team and its objectives and ultimately who who are your internal customers absolutely um thank you Alex and good morning everyone so I'm part of a central team of dbas and we manage several thousand uh databases globally distributed uh in support of all lines of business at Ford um in terms of data processing granularity uh they vary all the way from near real time and support of plan floor systems all the way to analytics and Mainframe um our database assets are distributed both on Prem as well as in the cloud with uh gcp being our primary cloud provider um in terms of uh database engines we we span the gamut all the way from extended got to be careful here extended relation uh to all all types of all elks of no SQL databases document processing graph databases wide column databases caching and so forth we're actually looking at time series and Ledger databases so we have a wide variety of uh database engines thousands of databases globally distributed as as I mentioned very sophisticated data granularity requirements uh in terms of our customers um we cater to all the personas that nimes mentioned um you know so so we uh whether we application developers Architects security Specialists executive management we need to make sure that our customers have the right Access Data as well as information so they can make informed decisions and get get their job done final comment that I would make is one of the core tenants of our database strategy at Ford is what we refer to as fit for purpose databases it's no longer one siiz fits all answer so that adds to the variation or to the variety as well as to the evolution of uh of offerings that we need to make available wonderful wow um so how how how exactly is your work so important to to to Ford as a whole um see we we consider ourselves doers of of of the data at Ford uh and and we see data as being a critical component of the compute fabric um and and and that allows us to build both Innovative applications as well as to introduce uh and integrate novel Technologies techniques such as Ai and then furthermore generative AI uh okay It ultimately uh you know leads forward to build uh great products and make better better and timely business decisions um so um you know that's that's sort of in a nutshell thank you uh so we know you've been migrating many assets from HPR to the cloud to deal with all the scale and complexity involved in in your database fleets what have been your two biggest challenges along this journey yeah two biggest challenges um that we fa as an organization not just the uh database organization but the larger Ford it organization number one we are involved in uh an aggressive multi-year effort to migrate large segments of our application portfolio from on Prem to the cloud to gcp uh so that involves U pretty much all the personas that n nimes mentioned and and you talked eloquently about the journey that's involved some applications have to be completely Rewritten some can be ported uh so so so there's a challenge and and we stand ready to support our partners within within the organization whether it's application developers um you know that require expertise or don't have the right expertise with respect to um the the new database offerings or feature functionality that exists on gcp or assist as part of the migration efforts uh troubleshooting uh performance issues and whatnot so we're we're we're Partners along the journey I think the second biggest challenge and I talk about this internally I'll I'll mention it here in our journey of two plus years three years to the cloud we've noticed a fivefold increase in terms of our databases um and and this is primarily a result of a new paradigm in terms of managing an operating ating databases in the cloud we refer to it internally as database as a service and what that means is the cloud provider pretty much handles all the operations that were traditionally done on Prem the relationship is a direct one between the application teams and the cloud provider uh so as such the role of the DBA and the DBA organization has changed and is evolving at Ford there's certain things that need to happen at the central DBA function level uh um especially around visibility Fleet Management um compliance and cost these are things that internally we refer to as uh driving a cloud mindset so all these things are things that are now uh handled at the at the the central DBA level this is a significant inflection point and I expect very large Global organizations are going through the same Journey we're going through so how have you dealt so far with with Central visibility I was hoping you were going to ask that question um so our approach we have built in very sophisticated internal dashboards uh that are Cloud specific that uh investigate the metadata that the cloud provider exposes by way of an API uh so these are time consuming U and a couple of things to keep in mind first of all the apis Evolve over time so you always have to be well versed in terms of of what the cloud provider makes available and second the data model uh or the metadata that's returned back to the the customer also varies in time so this is a a process that uh that takes time it's and it's not an easy one and it also results in the fact that the information that we provide back sometimes is disjointed obviously from Cloud to Cloud so our preference is to have the cloud provider do that on our behalf and I think we found an exceptional partnership with Google we've worked closely for over a year um you know providing our feedback in terms of our needs and as as we've seen Gemini and database and databases uh evolve we've we've been part of the journey and um and again an exceptional partnership meeting our needs uh both today and going forward so how how will Gemini and databases help Empower your team at present and how do you that evolving in the future okay um so in terms of the present as I mentioned we've been part of the journey so we believe that we believe it aligns with our vision uh it offers us the ability to provide Fleet level management to provide visibility across all all offerings that exist on gcp um you heard nimish talk about democratizing access to the data and I'll I'll I'll just add a bit of context to that what what we see that is uh meaning is that all the personas in a sense have uh visibility to the exact Fleet and potentially could interrogate and investigate the uh the fleet and take whatever actions are are deemed appropriate um so that working all right uh so that's that's part of the the the present now just thinking about the product itself I think it's both and I I'm going to be a bit provocative here I think it's both revolutionary and evolutionary so why revolutionary first of all this is the first product that Google is uh providing to the marketplace to handle the space uh and and from from my perspective and our perspective there's two really powerful things that it brings for customers first of all um you saw in the demos you saw in the pillars the integration potentially in time of these different um assistive capabilities whether it's in migration C- development compliance as well as Fleet level management and and view you don't have to jump around from one tool or to another that's a really powerful capability and then the second reason I think it's revolutionary because because it handles all the needs of the different personas that are mentioned um and and that gives uh the opportunity in very large organizations for the diff different personas to have uh continuous dialogue there could be issues raised at the CIO level or uh or someone managing a significant chunk of the fleet with respect to security issues compliance or cost or whatnot and they get circulated internally and all personas have visibility that's that's the essence of the the democratization uh and I think that's really powerful now why evolutionary so this is the the challenge that I put forth and put forth um we think that it it sets the right foundation going forward for what what I believe and we believe are going to be three major uh directions and you can consider them as inflection points or whatnot number one we believe that uh extending the cloud paradigm on Prem is going to happen especially in the database space and you'll uh you you there's presentation I think at this conference around the Google distributed Cloud um other Cloud providers are working on similar type of capabilities so that Paradigm and that way of operating managing databases I think is going to extend to on Prem so this sets the right foundation because it gives you the visibility the 360 view capability and whatnot second um it also sets a marker with respect to other database offerings that are made available through Marketplace because eventually they will have to integrate or be aware of what database or Gemini and databases has to offer because large organizations like cowers are going to want to have that visibility and then the third thing I think there's going to be C that's why I call it evolutionary capabilities for customers to bring in their their specific metadata that can be joined with did I complete with the metadata and the data model that's uh the underpinnings of Gemini in databases that's a long answer sorry about that that's a very good answer budon um we've learned so much today about your database transformation journey and really appreciate you sharing your thoughts also about how Gemini in databases can help you along the way um and over the past year and more it's been a huge pleasure to have partnered so closely with you as as a database leader and Ford as a driver of innovation so thank you thank you and I want to also thank the Google team I think it's been an exceptional Journey from our end so thank you thank you very [Applause] much thank you Ford for your partnership your thought leaders ship in shaping the future of database Management Enterprises like Ford TR Gemini because it's Enterprise ready from day one it's designed keeping data privacy and security needs of Enterprises you control your data not Google we do not use your data for any model training like all gcp products data access through gini is controlled by strict IM permissions we also have optimized have an AI optimized back end so that you get accurate and comprehensive responses we leverage our mature database infrastructure with industry-leading Vector searches last it's built on Gemini platform which follows Google's AI guiding principles and with Google's AI scalable infrastructure it can meet demands of thousands and hundreds of thousands of databases very easily like for you all can kick tires on Gemini databases today this all available to you right through our console I cannot wait to see what you all can do with Gemini in databases I want to welcome you all to the new era of database management thank you for attending the session today have a good day and rest of the next thank you so [Applause] [Music] much