 Hi everyone, I'm Stephen Tan, Chairman of Soda Foundation and VP and CTO Club Solution and FutureWay. With me today is a ready from Intel and together we are going to share with you what Soda Foundation is about and how we are trying to unify data storage management. The Soda Foundation focuses on open source data storage management. We launched the foundation back on June 29 of last year and this foundation was chartered under the Linus Foundation as a project. Also the mission that we have is to actually foster an ecosystem of open source data storage management software and also to provide a new term forum about different projects that come together to collaborate and integrate and ultimately this will help us to provide end users with quality end-to-end solutions. The Soda Foundation is supported by these members that you see here and besides the members we are partners like the Storage Networking Industry Association. So the governance for Soda Foundation is very typical of a Linus Foundation project. We have a governing board that obviously saw the business decisions and so on. And the governing board is supported by four committees, the technical oversight committee which actually drives the sets, the technology strategy and the roadmap. The end user advisory committee provides the use cases and feedbacks and the outreach committee is the one that organizes events and evangelizes the Soda Foundation and the Linus Committee focuses on collaboration with the different projects and fostering integration. So the people that are involved in the foundation is spread across the world and they represented this from different companies like big and multinationals companies as well as small startups. So for end users are very important to Soda Foundation, the foundation end user advisory committee as you can see here the representatives come from major corporations as well as smaller organizations. So we have like Toyota, Vodafone, South Bank and so on. So they represent some of the largest and most innovative companies around the world. And what they do is that they meet up regularly and they provide feedback on their use cases and also do a POC test and provide feedback on the test. And this helps us to help the TOC to determine the roadmap and set features for the roadmap and the directions for the Soda projects in general. So the ecosystem that we are trying to build comprises of our five different segments. So in terms of like the end users, these are the ones again that help us to provide us with the requirements, the use cases, the feedbacks and so on. The developers, the community developers are the ones that help us to develop the projects and provide innovation to Soda and the data storage industry. The vendors that are involved in these projects, I mean, they help to build the solutions and what's important here is that there are not only solutions, I mean, this will be open source solutions that are at the interoperate. And then the industry organizations that we work with like the standards bodies and other open source foundations or projects, we work with them on standardization certification co-marketing and events and stuff like that. And for academia segment, we just kind of started. So what we are looking for is like, you know, helping to set up labs and then do internships and have all this academic research on data storage technologies. So the Soda community, we've been growing it since the open SDS days, which was the previous life of Soda Foundation. So you can see like the photos that you see here, all the different events that we have held around the world pre-COVID and since then because of the pandemic that's going on, most of our events are virtual. So hopefully we can start having physical events again around the world by next year. Okay, so one of the things that we are currently doing is the data storage trends survey. This is a survey that we are doing in partnership with the Linux Foundation research team and also the partners for the survey they see a CNCS near OIF and so on. So we have kind of completed the survey back in just last month actually. So we are currently producing a report and hopefully we'll be able to share the report by the end of October. So if you are interested in the report, please follow us on our Twitter or LinkedIn and when the report is ready we will be announcing at those channels. So next I'm going to be introducing the open data framework. So this is the typical enterprise IT picture that you see. I mean you're like virtual desktops, you have like VMs, you have a file store backup development, you know, clusters and so on. So what's happening is that most of the time I mean you have like different solutions for each of these things and what happens is that each solution creates a silo that makes it hard to monitor and control. So the problem gets worse because you have all these different, I mean your data is spread out not only across these different silos but all, I mean it's also spread out across data centers, the clouds and the edge. So the key challenges that your phase is like how do you, you know, control the capacity or how do you optimize the capacity? How do you optimize the performance? How do you do data protection? How do you data security, data compliance and stuff like that. So having a modern data infrastructure adds to a different degree of data storage challenges. So what the open data framework allows you to do is that it provides a centralized view and control and connects storage to containers, VMs and other platforms. And this framework should provide data storage services such as a blog file and object storage, backup recovery, security and compliance and so on. So the whole idea is to unify data storage management with a single open framework across the core, the cloud and the edge. And here this is an example of how the open data framework can be applied. And in this picture what you see is like a connected card platform which is, I mean, which has a data center, the edge and also, I mean, it utilizes the cloud and also the warehouse for like archiving and so on. So with the open data framework, it can satisfy many of the requirements of this connected card platform by providing the blog file objects storage services, be it in the data center or at the edge or in the cloud. It does backup and recovery. You can backup from, for instance, the thick snapshots from the tier two storage and then back it up to the cloud or to the cloud storage. And also you can take care of like lifecycle management, for instance, data that sits on tier one storage that's been sitting there for idle for three months and we can actually move the data to tier two storage. And then after it's been sitting in tier two storage for another six months, we can archive it to cloud storage or move it to the cloud archive. Okay and there are, I mean, other things that you see here, the security and compliance and retention and archiving and so on. So essentially with this open data framework, everything can be done through a single platform and no matter where the data sits, whether it's at the edge, the data center or in the cloud. This diagram shows what ODF is about. ODF architecture provides an open API for integration with platforms and applications and offers like a seamless plugin integration with Kubernetes, OpenStack and VMware through plugins. The storage profiles offers policy-based storage provisioning, data protection, data life cycle and so on. And heterogeneous storage from the different vendors, they connect to the storage stock and using CSI, SwapFish and OpenStack Cinder, Manila interfaces and their performance can be monitored through a single pane of, through a single glass of pane. And the multi-cloud controller offers access to the different cloud providers through a single S3 interface. So it enables things like cloud tiering, cloud backup, and essentially it lets users easily build end-to-end solutions with any storage. So open data framework APIs are based on a SWAP standard. So by leveraging SwapFish management profiles, we're able to speed up development and ensure interoperability with the SwapFish ecosystem. The ODF storage API includes storage provisioning with policies for applications, laptops, retention and so on. The multi-cloud API supports the major cloud services and operations such as data life cycle management. For monitoring, we have the monitoring API that enables storage performance monitoring with IOPS, throughputs and other metrics. So all functions are with, I mean, the way we design this is that API is with API first design and so all functions are accessible and can be easily extended to support, I mean, newer features. So as I mentioned earlier, storage can be connected to ODF by the different management interfaces. Open Data Framework currently support all CSI drivers and allow multiple drivers to be blocking at the same time. Open Data Framework also supports OpenStack, Cinder and Manila. And also again, we support the SWAPFish standard. So with open source, open standard, open ecosystem and open collaboration, we think or we believe that Open Data Framework will be able to meet the needs of most end users and help end users unify data and storage management with a single open framework across the core of the cloud and the edge. So with that, that ends my part of the talk and then ready to be introducing the SOTA projects in detail. Thank you. Thanks Stephen for the SOTA intro and the Open Data Framework overview. I'm going to cover the SOTA incubation projects and the ECO projects. SOTA incubation projects follow our governance model that is actually provided by the SOTA foundation whereas the ECO projects follow their own governance model. But these projects are really meant to be providing a complementary storage and data management services for the core framework services. The first one in the SOTA incubation project is the DOs, DO stands for Distributor Asynchronous Object Storage. This is actually developed by Intel. It's an open source project. It is designed from the ground up to take advantage of the next generation media storage class memory as well as the NVV SSDs performance and bandwidth characteristics. This has all the attributes needed for the distributed object storage functionality. This includes the ability to make sure that the data placement looks at different part domains and distributes your data and it has the ability to provide the end-to-end data integrity as well as non-blocking IO operations. This includes metadata as well as the data related operations to deliver high throughput, high performance, IOs per second for the next generation media. The next incubation project is the IG. IG is actually contributed by Chana Unicom. This is another object store. It is actually being used in production by Chana Unicom. The concept behind this one is you can have multiple independent self-clusters. These are manageable without incurring the significant failure domain type of problems, especially when you have self-cluster that have multi-petabyte capacity for domain management becomes very challenging when there is a failure. IG addresses that gap by managing independent self-clusters and presenting a unified namespace using the S3 API, which is based off of Minio backend storage and it has its one metadata layer that uses the database in my SQL or IDB as a way to manage the metadata. That includes details about different self-clusters and where your data is located and security and other aspects that are actually maintained in the metadata. Those are the two incubation projects. The EcoProjects lint store is the block storage management software, open source software that is actually developed and managed by Linwit. The fundamental construct in the lint store is you can actually manage replicated volumes among pull-off servers and you can actually create these volumes from different types of media. If you look at the bottom of the right-hand side picture, lint store can actually carve out the block storage out of hard disk drives, the SSDs as well as NVMe media and includes persistent memory as well and it can expose the block storage via traditional ISCSI based on network protocol or it can actually expose to the NVMe fabrics, which is the protocol of choice for the NVMe devices and then you can manage these volumes natively using the container orchestration frameworks such as Kubernetes. Lint store does have several enterprise capabilities, things like multi-tier storage management. You can have a pool of storage managed for the hard disk drives and then you can have a pool of storage managed by the NVMe drives and you can actually migrate the data back and forth based on the life cycle management policies. Datadoop is another critical functionality specifically for high performance and expensive media like NVMe where Datadoop is an extremely critical capability that is required. Geoclustering where you can actually have multiple geos and you have the storage deployed in multiple geos, you can actually manage them together using the lint store geoclustering capability and then of course you want to be able to deliver an extremely high performing data access specifically for the NVMe type of media and deliver a significant number of IOPS with extremely low overhead data plane. So that's really the lint store. So it's a distributed block storage functionality is what you get out of lint store. The next one is open EBS. Open EBS is a container attached storage and it delivers that capability through the dynamic persistent volume implementation which is really required for stateful applications. Stateful applications like Cassandra, MongoDB, MySQL, they all require a way to protect your data beyond the storage server fault domain. Lint store, Open EBS actually provides that capability organically through the native integration with the Kubernetes container orchestration, container storage services. It has different data engines that you can actually take advantage of based on the type of media and the type of server that you have with a control plane that provides the driver functionality. Things like CSI drivers for Kubernetes have a way to manage and get dynamic persistent volume capability. So Open EBS is actually managed by Maya data. It's part of the CNCF sandbox project. Zenco is the other open source infrastructure software. This is the data management software stack, open source, provided by Scalety, where you can actually manage the islands of data that is actually deployed on-prem as well as public cloud. So you could have object stores on-prem as well as, let's say, have it on Amazon and Microsoft and Google Cloud. You can actually provision the object storage transparently through one uniform API which is basically based off of Amazon S3 and then you can provision them. You can actually manage them. You can move them all the lifecycle management aspects through the multi-cloud data management framework that Zenco provides. You can actually get to all the functionality needed. So that's on the Zenco. Cortex is the open source project developed by Seagate. It's a distributed object storage that is primarily designed to drive efficiency to take advantage of the high-capacity hard disk drives. It has a mechanism to integrate with the other object storage backends, things like Deus. So let's say Deus is managing the high performance in the media. You will be able to actually take a snapshot of that media and archive that into the Cortex object storage backend. So it provides a mechanism to coexist with the other object storage backends to deliver a complementary multi-tiered object storage solution as well. So those are the incubation and the Epo projects. So as Stephen mentioned, Soda Foundation and Open Data Framework, the focus is to essentially deliver unified storage management as well as the data management capability. All the services around it and then consume that using the cloud-native frameworks. The incubation projects as well as the Soda Epo projects are essentially meant to provide complementary services to deliver a comprehensive data and storage management functionality that the customers have been looking for. Whether it is delivering a high-performance block storage or high-performance object storage that can exploit the next generation media or it's a life cycle management or it's essentially high-capacity hard disk drives. Irrespective of what kind of storage management problem you may have, you will be able to actually manage your data and the storage through this integrated set of incubation and then Epo projects that are available through the Soda Foundation. We look forward and your help in joining the Soda Foundation. This may include anything from core contributions to the Soda Foundation services or contributions to incubation and Epo projects are providing the user perspective. Joining as a user are actually consuming the Soda and deploying that in your implementations as well. So, for sure, we can take advantage of anything and everything all the way from the development to requirements to testing and real life deployment usages. So, we look forward to your participation in the Soda community. We would love you to join the Soda Foundation. So, anyone who is interested, send an email to Soda Foundation and we'll be happy to have a discussion with you. Thank you. This concludes my session.