International Business Machines Corporation (IBM) Earnings Call Transcript & Summary

All right. We'll get started. So my name is Arjun. I'm a product manager here at IBM in Data and AI, and this is Ani, and he's a Senior Product Manage for data and AI here at IBM. And once again, thank you all for joining. You're all here because you're all super excited to learn about a brand-new approach to cloud data warehousing that we are offering with IBM Db2. But before we get into that, I want to take a couple of steps back and talk a little bit about Db2 itself. So for those who know and for some of you who may not know, we've been around for quite a while. Db2 has been out in the market for 30-plus years now. And in that period of time, we've been innovating starting with version 1 all the way back in 1993 and now making our way into the 2020s. And over that time, we've built up a tremendous amount of IP and built up a tremendous amount of trust with our customers and running mission-critical workloads for folks all over the world. And this shows up in our metrics and also in the kind of customers that we work with today. So not 7, not 8, not 9, but 10 out of 10 of the largest banks in the world run on Db2. And these mission-critical banking systems are there to support millions and trillions of transactions every day. So if you went to the store recently and bought a pack of gum from a convenience store there's a very good chance that, that transact ran through a Db2 database at some point. And we all know that the banking industry and the banking system heavily relies on trust and really, really value security. So if 10 out of 10 of the largest banks in the world, trust Db2 with their mission-critical workloads, we're probably doing something right in that area. And that's also the case in the retail industry as well. we're the #1 Fortune 10 retailer runs on Db2. So a lot of really, really interesting applications, and this also kind of carries over to the automotive industry. We're 9 out of 10 of the largest auto manufacturing companies doing Db2. But there's also some very interesting applications there. Some that we're all pretty familiar with in terms of predicting maintenance or freeing out waste to run operations safely and smoothly and also some new interesting applications evolving around autonomous vehicles and connected cars and all that good stuff. So whether it'd be in banking, auto, retail, that's just a small taste of our customers all over the world. And over the years, we've invested tremendously to ensure that they're -- that we are there to support all of our customers' mission-critical workloads, no matter what industry they're in and no matter where they happen to live. At the end of the day, that's our DNA. We've been in innovation for the last 30 years, and things have changed a lot from when we started to where we are now. And as a lot you are pretty familiar with today's data landscape, you know that today's data landscape is costly shifting and changing. And especially in the past decade, a lot of complexity has been introduced and Db2 has evolved to meet that complexity and service our customers with these new workloads. So nowadays, there's so much data that we have to be working with. And data coming in all of these different types of forms, and we're working with data applications and different data services and they're all growing. We also need to be able to work with this data in different ways. So we have things like transactional databases, but we also have analytical warehouses. We have a late cost engine. We have graft engines. We have no SQL engines. So there's a lot of different methods, a lot different engines to take this data and put it and work with it and get interesting outputs from it. Data is also just much more complex. It comes in many different forms and in many different formats so that's requires a lot of cleaning and a lot of attention before we can get it into the engine. And then lastly, in terms of storing that data, the infrastructure is available to store the data have become much more complex and flexible. In the past, maybe 20, 30 years ago, it was 99% on-prem. And now we have cloud and we have public cloud and multi-cloud and hybrid cloud strategies. So there's a lot of other options out there. And ultimately, we need a single engine that can handle all these different things. So if you're a data leader or a consumer, maybe a simple way to manage complexity with a single database engine built for any workload, any data type and any skill set. So if you are that data leader or if you are the consumer, and this is something that resonates with you, then you're in the right webinar because we've built Db2 as a single database engine that can handle anything from a data perspective, whether that be a workload data type of skill set. And we understand that customers today are aligning around a number of different strategic objectives. So whether that be modernizing global operations, there's a lot of challenges that come along with it. Downtime is a really important thing to consider, especially because it can directly impact financially your business. Applications need to scale faster and be very flexible. Imagine you're working in retail, the intensity of your data is going to be fluctuating a lot depending on the tech season. You can imagine you're going to have a lot more data to work with, in December around Christmas time versus now in March and April. And also the risk of embarrassing data breaches is really, really crucial to consider because it can be really costly when working with regulations and also it's just embarrassing to have that out there. And you can imagine these are some things that Db2 can help with, especially going back to the fact that we are we are in touch with the biggest banks in the world. So security is something that's on the top of mind. We have industry-leading SLAs. So we offer continuous availability. So downtown isn't something that you have to be worried about. Analytics as well is something to consider because being able to make decisions faster with incoming data has become super crucial, and it's become a really, really important need. However, when running these analytics that can become very, very costly. And a lot of times, you're working with data that isn't fresh and is not up to date. And so you need methods to be able to continuously update data. And so there are solutions out there that Db2 offers to address these strategic objectives and to be able to address these challenges. So for example, something that Ani is going to be touching on very soon is we're introducing a new generation Db2 warehouse that relies on cloud native object storage. So that is something that can really, really address concerns about the high cost of running analytics. Also, we offer capabilities around continual ingest, which will make sure that your data is fresh and that the dashboards that are powered by this data are up to date, which will allow you to pivot and make decisions faster. And lastly, you would want to be able to broaden the scope of your analytics so that you're not just looking at data in your warehouse, but also data in your data lake and be able to work with open data formats. And that is something that Ani will also be touching on very, very soon. And lastly, another strategic objective is something that's been a hot topic over the last couple of years in machine learning and AI. And we offer some capabilities there as well, running machine learning models right out of the box in Db2. So ultimately, we've built Db2 as a single database engine that can handle anything from a data perspective that you have as a requirement. And also, this is across hybrid, multi-cloud environments. And speaking of hybrid multi-cloud environments, cloud first is the new strategy for Db2. So IBM Db2 has now moved to a cloud for a strategy that means that all of our features and releases will be coming to cloud first, and that is the highest priority. And so why that's really important is all of our innovation is focused on that. And it will be deployed a lot of our features will be -- will deploy fully managed services to IBM Cloud as the primary cloud of choice. However, we're also engaging opportunities to bring all of our products that you guys all love or are interested in to where you are. So for example, this I think to be a lot of interest. IBM has signed a historic collaboration and strategic agreement with AWS to deliver IBM SaaS on AWS. And we know many customers prefer using AWS as a cloud provider, and we want to bring Db2 to them where they are most comfortable. And in fact, Db2 Warehouse was the first AWS service that IBM launched almost 5 years ago and that leads perfectly into a lot of really cool things that Ani is going to talk about regarding Gen 3 and why you guys are all here today.

Perfect. Thank you so much, Arjun, for that amazing introduction. Db2 is definitely 1 of the most critical pieces of infrastructure that's running the world today. Now I want to kind of pivot away from the product Db2 into and towards Db2 Warehouse and specifically the fully managed versions that we offer. So Db2 Warehouses are flagship cloud data warehouse that is fully managed -- that's a fully managed elastic cloud data warehouse that's available on IBM Cloud and AWS. It runs on -- it runs on the Db2 engine but that's customized and optimized for us. Things like we do a large portion of processing in memory so the data is pulled from storage into memory, and then we do all the processing there. So you're getting some really snappy performance out of Db2 Warehouse. We're also able to do things like query in compress data sets. So your data that resides in storage is typically compressed at like 4x, some data types, slightly more than that. We're able to run queries on those compressed data sets so that its saves time and cost of decompressing the data and then processing it. And finally, we also feature data skipping, which means the engine has the ability to intelligently identify which data would be needed and only pull that data into memory, the data that is required to satisfy the query and leave everything on storage. This massively improves performance and while reducing the compute overhead. Also, we do support a ton of other capabilities that Db2 has been known for, stuff like industry-leading SQL compliance, things related to security and encryption, in database machine learning, machine learning query optimizers and so on, the list is really long of the capabilities that Db2 provides. Third, I want to talk about scaling. So it's a scalable and elastic solution and compute and storage are separate from each other. So if you need more computer storage, you can use the UI or leverage our rest APIs to scale them up or down. What we see our customers do is that they scale the compute down on the weekend or days when they don't need a lot of compute power and scale up during periods of high requirements, say, let's say, during the holiday period. We also see customers leverage this capability to perform periodic activities such as ETL, where they need high compute for a small amount of time, and then they scale the data warehouse back to its normal configuration once it's done. And because customers pay only for the compute that they use, they can cost optimize very effectively without having to worry about wastage of money. Fourth, I'm going to talk about continuous availability. So Db2 Warehouse is continuously available. And what that means is we can think about it in 2 separate components, right? So let's talk about compute. So from a compute perspective, the entire deployment of Db2 Warehouse is containerized. So we take advantage of container platform capabilities to do failure detection, failure recovery in place. If one of the nodes in your cluster happens to go down, the Db2 Warehouse will detect that the node is in a failed state and will take that node out, swap in a healthy node and all of this happens in place without the customer having to spin up a new cluster and restore data that has been backed up. None of that needs to be done because Db2 is able to kind of fail and restore in place. From a storage perspective, today, the data resides in hyper farming, cloud-managed SSD, block storage, that is fully managed, highly redundant and highly available. Customers that require backup and they require disaster recovery, you can offload back up to a different region. If an entire region, happens to go down, you can spin up a new cluster in the new region, restore your data into the new cluster and you will be back and running within a short amount of time. So we've done a lot of work to make sure that we can handle any kind of disaster scenario. And finally, I want to talk about reliability. We have comprehensive self-service backup and restore capabilities along with automated scheduled backups. So the customers can decide when they want their data to be backed up and with what frequency in case of disaster customers can quickly restore from -- restore their data from their snapshot backups. This backup file is replicated across different availability zones and regions, which gives customers the peace of mind that their data is safe no matter what. So Db2 Warehouse is currently available across the world on IBM Cloud as well as AWS. We have customers from all over the world who take advantage of Db2 Warehouses rich feature set and high-performance capabilities. So now I want kind of talk about like a select set of our customers who are using Db2 Warehouse today to achieve their analytical objectives. We've got companies such as Marriott who are using Db2 Warehouse to build personalization models based on their customers' preferences and behavior so that they can better serve their customers and provide them with better service. We've got companies like Active and Valor using Db2 Warehouse for price optimization, sales analytics. Db2 Warehouse is a perfect solution for companies that are looking for -- looking to perform real-time analytics, building predictive models using machine learning building reports and dashboards to enable business intelligence and also securely share data in a [indiscernible] way. So hopefully, I've been able to kind of put forward the point that Db2 Warehouse is already a super great data warehousing system that can handle all your analytics needs. But now I want to kind of talk about some of the new things that we are bringing to Db2 Warehouse that we've been working towards for almost 1.5 years. This is the biggest update we've brought to Db2 Warehouse in a long time. And today, we're going to be introducing some truly powerful features for our customers. First and foremost, we are launching support for our cloud optic storage with Amazon S3. This will allow customers to land Db2 data inside an S3 bucket and get the economic benefits of running on cheap object storage, petabyte scale object storage. We have built a proprietary cache and capability that is -- that intelligently caches data into the NVMe cache, which is local to the compute. This enables faster and better performance while running your workloads, and at the same time, storing data on object storage. We are pretty excited about this new development because this will fundamentally change the economics of running the Db2 Warehouse service. We have some preliminary numbers, performance numbers that we are seeing in the lab that we want to share with you today with this new cache in capability. The second feature that we are launching is the ability to interface with open date formats. Arjun alluded to this a little bit earlier. As we all know, the industry is moving away from prior data formats and customers are looking towards Open Data formats such as Iceberg, Parquet and AVRO to store and transport their data. With the launch of the next-generation Db2 Warehouse. Db2 Warehouse will gain the ability to interface with the data that is stored in open data formats, import that data into Db2 Warehouse, run queries on it, export the data into an open data format and share it with trusted partners and ecosystems through these formats. By the way, the open date formats form the foundation of IBM's data lake solutions. Through this release, Db2 Warehouse will become an integral part of IBM's data lake solutions by allowing metadata and -- or data to be shared through catalogs and meta stores with trusted entities. Finally, we know that our customers are looking to run truly large-scale analytics workloads that require a very high amount of compute. So we are expanding the compute scaling of Db2 Warehouse from 576 cores to almost 3,000 cores, that's almost 6x improvement and jump in the compute scaling. This will allow enterprise customers to skill their warehouse as much as they need without worrying that they might run out of compute and also have combined this with this petabyte scale object storage, Db2 Warehouse is ready to handle pretty much any analytical workload that can be thrown at it. Apart from these key features and there's also a slew of other features that we are introducing through the next generation, which we call Gen 3, the third generation. Features such as granular backup and restore, which allows users to back up specific schemes and tables instead of having to take a backup of the entire system. We're also introducing new integration capabilities through IBM App ID, and also, we are launching new APIs that will allow users to perform all these operations through rest APIs as well as the UI. So this was like an overview of all the features that we are coming out with in Gen 3. But what I'm going to do now is I want to kind of dig into the 2 most important ones. One is the cloud object storage and the other is open data format. And what we've done is we want to kind of show you a small demo of the cloud object storage working in real time. This is happening on a cluster that is currently live on AWS. Here, we can see the console. What we're going to have is, we're going to take a representative workload of -- representative analytical workload, and we're going to run on that both on block storage as well as optics storage to see how they both perform. So we're going to head on to the SQL editor. The dashboard right now, we're going to head on to the SQL editor. So like I said, what we're going to do is we're going to perform a basic cloud operation where we will create a table. We will insert a large set -- large data set into it, somewhere around 14 billion to 15 billion rows of data. Then we'll perform a SPSS update, which is basically updating one column for 1% of the data, and then we'll commit this data back into the database. So right now, we are doing this on block storage, which is the current storage system. This is to set a baseline of performance to see how much time it takes for us to execute. And then we'll perform the same operation on object storage. So once we trigger the operation, we're going to speed up in third part. It took around 202 seconds to insert data for a total of 210 seconds for this whole operation to be performed. This sets our baseline for Gen 2 or like the current generation, which already uses high-performance SSD, so it's by no means slow at all. But now we're going to trigger the exact same sequence of steps on S3, which is the new caching engine, and we'll see how much difference we see in performance. So the same steps, again, we're going to create a table. We're going to insert 14.5 billion rows into it. We perform a SPSS update and then will commit the data into the database. So as you can see here, I think we've completed the insert and it took around 68 seconds as compared to 202 seconds for Gen 2, and if you look at the overall numbers, we are seeing the entire operation was done in 78 seconds instead of the 210 seconds that it took for Gen 2. If we just dig into some of the preliminary performance numbers, we are seeing that Db2 Warehouse Gen 3 is about 2.6x faster in select performance, 2.8x faster and insert performance, a whopping 7.2x faster than update performance. And all this while using storage that costs 23 terabytes per month instead of 874 terabytes per month, which is what lost storage cost today. So it's an almost 38x drop in costs for storage while seeing a significant improvement in performance. Now 1 caveat that I would like to point out here is, is that being performed in the lab with a test workload with our best understanding of what a representative analytic workload would look like. Active numbers input production environments may vary a bit, but we will be putting out more comprehensive performance numbers in the coming weeks and months leading up to the launch. Secondly, I want to talk about using data lake tables, using open data formats. This is the second most important feature that we are launching. And what we're going to be doing here is we are going to be taking 1 table that is in Db2 format, and we're going to intersperse that with a table that is in a data lake table, which is an open data format, and we're going to run a query that joins both those tables together. By doing this you can take advantage of both open data formats as well as the performance characteristics of Db2 engine. So this is the data lake table that we see. Now we're going to head on to the SQL editor. We're going to run this query, which joins data from the open data format table, which is the historical weather data and a query that joins with the Db2 table format. We can also export this data out from Db2 into an open data format, which can then be used to share data with other trusted participants. And this data will be stored in S3. As you can see here, for example, this table has been exported as an Iceberg Parquet file into an S3 bucket, which can then be accessed by the user and shared with other entities that they want to share this data with. I want to talk a little bit about the journey from Db2 Warehouse, the Data Lake and back. So customers that today have analytics appliances such as IIS or SaleFish or PBOA systems, they can modernize their Db2 Warehouse is tailor-made for such migrations. You can update your existing appliance to the cloud using Db2 Warehouse. You can also replicate data from your mainframe systems to Db2 warehouse using light winning. Third, you can cost optimize your storage by choosing essentially like a combination of block storage or object storage based on your requirements, you can take advantage of the highly reliable cheap petabyte scale object storage to lower your storage costs. And fourthly, you can move data, share data with your warehouse data with the data lake by essentially using the functionality that we looked at by exporting data into open data formats. And also, you can promote data from the data lake into the warehouse. So you can essentially stitch your entire analytics environment together to get like a single entity, which can query data across your Db2 environments, your appliances, your mainframe systems, you have open data formats. And this is the first time that Db2 have had has this capability. That is pretty exciting. So yes, so this was like a brief introduction to all the new amazing features that we are launching in Db2 Warehouse Gen 3. We are launching in beta on April 17, and we're going to be generally available sometime in Q3 on AWS. Like Arjun mentioned, AWS is an important strategic partner for us. We have a strategic agreement with them. And we -- and Db2 Warehouse has been available on AWS for almost 4 or 5 years, and we will be bringing this new version to AWS, simply followed by IBM Cloud in the coming quarters. So we hope that you enjoyed this presentation, and we hope that you will sign up for the beta. The beta links are available. And now we would like to move on to any questions that you may have, and we'll try our best to provide you as much information as we can. Thank you so much for joining the webinar, and we'll now move to the question-and-answer session.

Hello, everybody. Thank you for the webinar. We've got a bunch of questions that we've tried to answer during the webinar, but then there's a couple of others that we can see. I'll just go through some of the ones that have been asked repeatedly. With regards to performance, first of all, the demo that we had is -- was based on a lab setting for a workload that we think is representative of our workload, but we will be publishing more comprehensive benchmarks in the coming months, where we will try to provide more complex scenarios and test out the performance of the new Db2 Warehouse Gen 3 engine and caching during that. I just going to quickly go through some of the questions that we have here. The first question is, what do you mean by replicating or even by replicate data from Db2 Z to Db2 Warehouse with light printing? That is basically capability that we have built through a collaboration between the main frame team and the Db2 team where you can move data that is on the mainframe and replicate it on to Db2 set of the data that you want to probably interface with other applications. That is what we mean by the replication from Db2 Z to Warehouse. The other question that we have is what is the migration path from Gen 2 to Gen 3? So currently Gen 2 is both on IBM Cloud as well as on AWS. There is a very straight forward path for customers who want to migrate from Gen 2 AWS to Gen 3 AWS. We will be providing all the tools necessary to easily migrate. The IBM Cloud to AWS is a slightly more complicated activity, but we will be launching Gen 3 on IBM Cloud very soon as well. So unless there is a specific reason that to move cloud providers that there will not be any difference in capabilities in the coming quarters. The next question that we have here is the beta for, yes, the beta for Gen 3 will be available from April 17. We're going to have 2 groups of data. So we're going to have -- the beta is going to run from April 17 to almost the GA period. So lots of opportunity to take advantage of that. The next question we have got is Db2 to data lake is through data virtualization. So this is an important question. So I would like to kind of elaborate on that. So the data lake that we have is based on the Open Data format such as Iceberg, Parquet, AVRO, ORC, et cetera. Db2 is through Gen 3, we are adding native capability for Db2 Warehouse to read and import and export files that are based on the open data format. So Db2 Warehouse will not need to take advantage of other IBM that help you to do this for example, Data Virtualization Federation. This will be a native capability. It is a pretty important 1 because what that allows us to do is it allows us to give customers a one-stop shop solution for all of their data types, so they can have their Db2 data that is already in their systems in binary format and then they can also interface with their open data formats thus kind of integrating Db2 Warehouse into their overall data lake solution. The next question we've got here is, is there any change to the Db2 licensing model given that huge scale out on core capability? That's an important question and there is only limited information that I can talk about how the licensing and pricing is going to work. But we will have more information for everybody as the months go by. We will have all the details available by the time we launch the product in May. The next question we've got is what Db2 Z to Db2 Warehouse replication software is used? Can we replicate Db2 Z data to S3 Parquet? Now that is a 2 step process here. From what I understand, although length to kind of determine what the path should look like. But overall, if we talk about it, it would be something like move data from Db2 Z to Db2 Warehouse through light printing and then exporting it out to S3 Parquet. Both of these are native functionality and they can both be done through start procedures and rest APIs. Therefore, it should not be a very complicated step. Although, like I said, the actual use case probably needs to have more in-depth analysis and more specifics before we can make a conclusion. What else do we have? What is the name of the tool, which can help in light printing? I can see that there's a lot of interest in the Db2 Z to Db2 Warehouse migration. What we will do is we can public a proper -- some asset that can talk about this at length. I think that will be useful for the people who are interested in that. How about migrations from IIAS, big volumes to Db2 warehouse on cloud, how do you address that assuming short time for switch over? That's a very important question. So Db2 Warehouse Gen 3 is tailormade for IIAS migrations. We are building a lot of consideration into the product when we are thinking about IIAS customers. We want to give them the easiest path possible to get their data from their appliance onto the cloud system and take advantage of Db2 Warehouse. The actual details around that would be released in the coming months. And we will have -- we also have IBM Expert Labs doing this migration available for customers to kind of do this. We are working very closely with IBM Expert Labs to help them understand what are the best possible paths to migration. And we hope that customers will be able to take advantage of whatever native functionality we built in to Db2 Warehouse as well as what we can provide as support, both from like a document perspective and from a live consulting engagement perspective to help them migrate. The next question we've got is what is the fastest way to bring in IBM Db2 warehouse in the cloud and how long it takes? So this is another important question. It is difficult for me to give a specific number of how long it takes. This is 1 of the product lines that we want to provide a very streamlined path to cloud migration. So you can expect some sort of tooling and some sort of streamlined process that's been defined to move data from on-premise to the cloud. We can have a longer conversation about that if that is something you are interested in. We can reach out to you and then we can have that discussion. Can we also process data from Azure and GCP storage buckets? We are supporting Amazon S3. That is the primary storage bucket that we are supporting but we know that Azure and GCP product storage is also very popular with customers. So we do have it on our long-term road map to cover at least a subset of the other storage options that are available in the market. It's hard to provide a very specific timeline, but rest assured it is on the road map. And we want to give our customers as much flexibility as possible in terms of where their data can reside while they take advantage of Db2 warehouse. One thing that I would like to say, which is not specifically connected to the SaaS version of Db2 warehouse, but we do have reference architectures of Db2 Warehouse that can be deployed on Azure today. So if customers are interested using Db2 Warehouse on Azure, they can do that, although it's not a SaaS deployment, it could be a self-managed cloud deployment for the time being. Next question is you said import of open table formats, what you really mean accessing Iceberg, Parquet, AVRO in place just as accessing external tables? So it's slightly different from the external table capability that we have today. Basically, what we are doing is we are kind of bringing the capability of reading and querying and ingesting data into Db2 Warehouse. So it would not be just a virtualized layer that will be queried. You can actually import data that is in Parquet Iceberg format into Db2 Warehouse and you can export it out as well. It's not just like a materialized query or a virtualized query. The functionality that we were trying to -- we are trying to market in this because that is what gives us the ability to synchronize with the data lake and export data out to the data lake, if there's a query, if there's some data that needs to be externalized or shared with other partners, ecosystems, products while at the same time if they are specific query workloads that require really high performance capabilities. That data set can be imported into Db2 Warehouse where you will get the same performance characteristics that you get today, which is what we wanted to do with Db2 Warehouse Gen 3. So the next question we've got is, is it based on [indiscernible]? No, Db2 warehouse runs on an analytical focused version of the Db2 engine. Like Arjun mentioned earlier, the Db2 engine has been around for a long time, and we've really in almost 30 years of innovation to add the performance characteristics can be as good as possible. The Db2 engine is probably one of the most performant engines in the market. And Db2 Warehouse takes advantage of that. So yes. So the next question we've got is when Db2 is storing this data on AWS S3 or IBM [indiscernible], do we still have the same structure data made by containers via the faster on cost than block storage. So I provided a bit of context on that during the video, but I can kind of repeat that. Basically, the idea is that with Gen 3 what are doing is, we are moving from block storage to object storage. But as we all know, block storage is significantly faster in its native format than object storage, but it is significantly more expensive as well. Which is keep improved performance while reducing costs, and the way we are doing that is we are using NVMe cache that is local to the compute that is being deployed. So what we do is we've built a caching layer on top of the Db2 engine, where when a query is fired the engine will first check if that data exists in the cache. And if it does not, then -- only then the data will be pulled into the cache and then processed. The Db2 engine is anyway very good at intelligent the Db2 engine for a long time. But with this cache, it gives like an even higher level of caching capability. That is how we achieve faster performance on cloud object storage than on block storage. One thing to keep in mind is that the cloud object storage that we saw in the video was a warm cache, which means it has been updated with a portion of the data which is what you would expect in a normal production environment as well. It's not a special scenario. But that is the way that we get better performance on cost than on block storage. The next question we've got is when can we expect Gen 3 on-prem running on OpenShift? That is another 1 of our strategically important areas. We want to make sure that these capabilities will be provided to our on-prem customers as well, where they can take advantage of object storage, that they have stored on-prem and reduced their cost. We do have a road map for the on-prem deployment as well. But like we mentioned earlier, we have now moved to a cloud first strategy which means all the new features will be first released on the cloud, and then they'll be brought to the on-prem versions in the coming quarters. So rest assured we will be bringing these capabilities to on-prem and they will be running on OpenShift as well as other popular Kubernetes platforms. So that we can give customers as much breadth of choice that we can to deploy Db2 Warehouse. The next question we have got is S3 storage provides eventual consistency models for delete updates for keys. Do you recommend S3 storage for real-time BI work workloads? That's a very important question. The answer to that is, it depends, right? When you say real-time BI workloads that term is very, very broad. So we can have conversation about the kind of latency and the kind of performance that you are looking for. But what I can say as a generic statement is that we are making sure that the real-time workloads such as dashboards and business intelligence are related. Workload should be supported through even the new engine. You will not see a significant difference in what you have today versus what you will get in Gen 3, the performance will be better, will be equal or better, I would say. But again, we are happy to kind of have a longer conversation about your specific workload requirements and discussions about how we can help you size your workloads or understand how they will perform. All right. So I think I've covered most of the questions that we've got. I can just go back and look at some of the answered -- questions that we answered during the call. if there are any in the interest of providing details. We've got a few questions related to performance benchmarks. And what I can say is that we will be releasing more comprehensive benchmarks on the performance of the new engine, which will give much more clarity about what are the performance characteristics for various use cases. This is something that IBM has been doing for a long time. So we will continue to do that. The beta version links will be available. I think I'm not sure if it's been put in the chat, but if not, we can always provide those links through e-mail or through follow-ups. Please sign up for the beta, we would hope that you are able to provide us quite valuable feedback and while taking advantage of new features and understand how the new engine performs. What else do we have. There was a question about snapshot backups. So the question was, while taking snapshot backup, it also takes backup of the Db2 Warehouse and Db and in turn, will it cause a downtime during the time of snapshot backups? This is an important question because it kind of talks about the backup and research capability. What happens is Db2 Warehouse will pause the right, so it will go into a right suspend mode from a technical perspective. And then we'll take the backup, the backup takes around 5 to 10 minutes based on the size of the volume and then it will basically resume right. The system will be available for reads for everything else for administration. But the rights will be paused because we're essentially taking the backup of the storage system. What else. We are on the question related to the parallel queries, which is again more specific workload characteristics that we'll be talking about during the benchmarks that we released. Let me see if we've got any other questions. Does it make any difference if it is AWS public or private cloud? So again, this is a nuanced question. We do have capability to connect with virtual private clouds inside AWS, but there are specific cloud types of AWS such as government, et cetera, which requires slightly more discussion about whether it will be supported. So if there's a very specific type of AWS cloud type that you want to have a discussion on, we are happy to have a discussion with you about that and understand your needs. Is there any -- are there any other questions that I can answer, please feel free to put them in the chat? We'll be -- we'll stay around for some time answering any more questions. And you can always contact us in -- and a follow-up on any questions that have been asked today or if there are any other questions, we are always available to answer your questions, both technical or from a migration product, whatever. I think there aren't any other questions. Okay, we've got one. Light printing, is it different to IIBR/CIBC? Like i said, we can have a conversation about this, the light printing question, we will reach out to you and then we can have a discussion about that. For smooth migration, do we have data pipelines? That's an important question. So Db2 Warehouse obviously works very well with IBM's flagship ETL tool, which is data stage. And the Db2 Warehouse also supports all kinds of API and SQL-based operations. So it is very integratable into any kind of existing data pipeline. We also have partnerships with Informatica for other ETL solutions. So yes, we do have data pipelines. Now for specific migrations between, let's say, Db2 warehouse software to SaaS, versus or IIAS to Db2 Warehouse SaaS, we will be providing specific capabilities so that you don't have to set up a lot of external pipelines. It will be very easy to do the migration. Yes. Db2 data lake is one of the solutions that can be used to replicate from Db2 z/OS to Db2-LUW or Db2 Warehouse SaaS. But again, questions related to Db2 Z are a bit more specific, which can be answered on a different forum, we will stay in contact with you to get those answers. All right. I think we are coming to the end of the presentation. Please send us any other questions if you have. And we can continue to have a conversation about that.