 Good morning and hello everybody and welcome to our today webinar and today our topic is optimizing business application when working on CRE-ODS. We'll be starting with just a quick glance at what is CRE-ODS and what is the CRE-ODS brings to the AIRT observation project and then we'll talk about the possibility of cooperation with CRE-ODS and with Cloud Ferro regarding your data, your processor, your application. Then I will tell you how to achieve the business sustainability for your projects and then we'll quickly cover some basics about the Cloud billing models and how to use them to your advantage. So currently CRE-ODS is the best platform on the market for the AIRT observation related projects. Our platform is focused and orientated for the providing the data and providing the cloud environment. So we are very focused on availability on the data, on accessibility on the data to have all the data available in real time and not waiting for any kind of cash or any kind of long-term archive that you have to access and wait. We provide very competitive cloud environment with many different versions of and flavors of virtual machines and also many other services. We are totally European company with all our public clouds hosted in European data centers with high environmental and safety standards to ensure that the business is environmental friendly and that all operations are safe. We provide all our users with free data, free Copernicus data and also other your data archives with many options how to access the data and how to connect to the data from our cloud and also from external platforms. We offer our users an option to incorporate their own data into the archive and for the users inside the working from our cloud the access to the data is always free and we try to provide the access as fast as it is possible. So for most of the clients or for all of the clients there are no obstacles in accessing the data in the speed of the access of the data and they just can focus on data processing. So CRE-ADS is not only a platform for providing the data and for data processing but it's also a platform for cooperation so our users have options to deploy their own applications, they own services, they own data and they own processors into the CRE-ADS and we can help them by sharing our experience and also by sharing our tools to explore putting their data into inside our own archive or creating private archives for the users to share data between many different projects. This is something that is very keen to us to see and to enable our users to share the data and to share the knowledge and to share the processors to better understand the data. The most common grounds on which we are cooperating with the clients and with our users is sharing the data especially our data between different users and putting between different collections, putting free or paid processors inside our tool chains or just selling the applications. The common frameworks for corporations that are ongoing on many different levels even right now is hosting the data. So if we see the business case for hosting additional data on our public repository we'll put it there and we are always open to suggestions from users what kind of data they are lacking and what we should put into our repository for them to process. We are also able to host any collection, any paid collection on our services and back it up to our experience and our user base and sell the data to users. So we can, for example, you can put your collection of data on CREODYS and sell it to other customers and we also provide customers with access to external data. For example, in this way you provide access to high-resolution satellite imaginaries. Images when users are able to order and access higher resolution data from external providers but they are essentially accessing it in the same way they are accessing the data on CREODYS. We just have to wait for us to transfer the data from the provider and they can also order the data, order the high resolution images to be taken in the future from our platform. The second and I think the most interesting way to cooperate is to host your data processor in our pipeline, in our chain. So this can basically be done in two ways. Either we can just add your processor as the next processor on our list and then you and all other users can use it to process the data and we charge everyone a fee that is covering the cost of running the processor or you can, we can put the commercial processor on the CREODYS and then we share with you the revenue for using the processors. So basically part of the revenue goes to the owner of the processor and part of it goes to us. Second option is I think this is the one that has some future because the processors are starting to be really complicated and we start to see the complicated chains of processing and for many users when there is just, there is no really need to set up all the chains. It's easier just to order the processing with some parameters and get the results and not have to not have to create all the environment. The third option of cooperation is just using the, just us selling commercial licenses for the software for DS users. So for example, if you have a software that has a license and you sell it to the users, we also work as a marketplace to sell the license to the end customers. And I think as the biggest platform for observation or one of the biggest, we are the place that many users will find it easier to buy the license here and not on external somewhere else to use it fast and freely. And on top of this, if your processor or your data or your license becomes part of the Crayodias, our support will support the end users in using this data processor or application on their VMs and will help them install it and use it. So this part of helping the users and the customer support will then go to us and we'll have to do this now too. So this very, very fast and very easy way to just give users your data and your software. Project business sustainability. The first thing I want to tell you is that for ad observation projects that are just starting or just testing or just searching for users and customers using Crayodias is one of the easiest way to to get the attention and to start working with end customers. So Crayodias is stable, Crayodias is quite big, so there are no infrastructure boundaries. Crayodias offers already rich collections of data and many other processes and users don't have to search for them in other places. It's a lot a lot of easier to run it on Crayodias than to set it on your own and try to create your own environment for the users who don't have to get the data and get other elements of the process chains to work. And Crayodias is very easy to develop, test and deploy applications. So you can start your environment in a matter of seconds and just start doing things and don't wait for it to pop up into existence and don't wait for everything that is always related to starting new projects. And if you start on the Crayodias application, it's costed on Crayodias, you can get our support to help you with your application. So for example, our support will support your users in many different ways and will be the first and second line of the support. So our support is currently done in three languages and we can add more if there's a requirement. We can do together some kind of promotion. So we will tell our users about your application, about your data, about your processor and how to use it. We can do together a webinar on how to use application and how to show what will be the benefits and how does it shine comparing to all the other things. And by doing so we can help you grow because the best way to have business sustainability is to have users that pay for the application and use the data. And frankly, most of the clients that your clients for your application are clients that are already on Crayodias. So they have an account here, they do some computations on Crayodias and they are familiar with the whole environment. So when you are using Crayodias, there are basic business models that you should be aware of. They are common in the cloud world, but you just have to know them to get a feeling. When we are talking about the cloud business model, no matter it is a VM data access or processor, we mostly talk about the per usage model. So the customer pays for the using, not for the possibility of using and not for not doing anything. So basically most of our customers run VMs and start and shut down dual machines or network on daily basis. When they need, for example, start 100 VMs, they do the computations and they shut them down to reduce the costs of the VMs. And this is perfectly normal and this is how we should think about cloud in general and about cloud services. So there is no need to have this big chain that is active all the time and only for some time you are using it and really processing the data. And extension of it is that you don't have your own environment, but you only order some processing of the data and you get results put into your storage that you are then transferring somewhere else. This is the cloud 2.0, the next model that is that you are only paying for the function or for computation or for the data, but you are not really running any calculations by yourself on your own virtual machines. And this is something that we provide today. And we think this is the way that market be moving in the future. So there is no need for the customer to be very experienced with Kubernetes virtual machines. The customer and the user should be experienced in data, what type of data, how to change their data, what is the enter of the results and how to use it. The end result data not doing the computation and running complicated computing environment. And the second model that you, the model that is more, let's say from the old days when how it was done in the old times, it is the fixture model. So in fixture models, you rent specific resources, computing resources or storage resources for specific time. You then, regardless that you are using them or not, you are paying for those resources. For us, it's more from business perspective, from cloud provider, we know that you will be using the computing environment for a long time. So the price is fixed and the price is much cheaper than, for example, one day costs 40% less than in pay use mode, but it's, well, you have to use it. Regals that you are using it or not, you have to pay for it. So you pay also for the dry runs and for times that you don't use the virtual machines. And you are committed for the long time for paying for the resources. So this, it reduces flexibility, but if you know that you'll be, for example, processing the data for six months, it will be much more easier to use the fixed billing mode and not the pay use mode. It depends on your computation and on your, and what you want to do and how long it will take. The per use mode has two distinctive modes. One is the pure per-pay use mode when you just upfront buy some resource unit from us. And then when you are using a resource, it has a price in resource unit and you have to pay for, pay for the VMs or for the storage in those resource units. And this is the basic, very, very quick model to run and very quick model to set up. You can go on the Creatia's website and set up an account, pay and start using in, I think two minutes is way too long to do it. So in one minute, you will be able to start your environment and start using it. But this model also has some disadvantages. The biggest one is that the resource is not guaranteed. So when we look at the cloud environment, the size of the cloud is limited, no matter what we try to say, the size of the cloud is limited. And because of this, there is no way for us to satisfy all needs. So because of this, you have to remember if you are running in pay per use mode, you can just run after, we can just run after the resources and won't be able to start more virtual machines or more storage or other things like this. And there is a unknown cost of the resources. So by unknown cost, I mean that you pay, for example, as 1000 euros and 1000 euros, let's you run five VMs for three months. But it also can let you run 1000 VMs for 15 minutes. And we don't limit you in any way. So you have to control your costs because you can easily spend everything and don't do much computations, if you have, for example, an error in the code. There are some fail sets there, but basically you are free to use all our infrastructures. And there is always a possibility of using it way too much that you really need. And there is a second building mode, the pay use mode is pay as you go mode. In this mode, you sign a contract with CloudFaro, paper contract. And then we let you use our environments or other products or processing freely. And at the end of the month, we invoice you and how much you really used during this month. So you don't pay upfront, you pay after using cover services. And again, if there are no fail saves, we can put fail saves here, but by default, there aren't any. So you can use it freely, but freely, it is quite easy to spend 10,000 euros per month for resources. So you have to control your usage. And those are the disadvantages. But because there is a written agreement, we can put some fail saves there. And if you want to be invoiced more, that you stated in the agreement. So how to let's say exploit those differences in the building modes and exploit them to your advantages. So the biggest advantage that you can exploit is the one that you don't have to pay for the resources that you stopped using in the pay use modes. So you have to analyze your project and be certain that you, if you need a VM running 24 seven, you will only need it for four hours a day. If you want to need it for four hours a day, it is easier to set up a pay per use account or pay as you go agreement and work there. And the best way to launch your environment in fixed time, you're really ranging it from the e-commerce side. So it takes some time and it's a long process because you have to pay for the invoice. And especially in pay as you go, you can just start and stop everything from the API. So for big environments in thousands of VMs, it is easier to run it that way. There is also a possibility of using the mixed billing mode. So for example, by mixed billing mode, I mean that part of your infrastructure or part of your environment is in fixed time and part is in mixed mode. There are three product categories that are way easier to to be managed in pay per use mode or in as you go mode because it is way, it is just uncertain at the start, how much resource do you need? It is the product generation when it's sometimes hard to determine at the start how many products will be generated and putting in a fixed agreement. The reality is that you will always order too much. The storage when it is, when you are having a fixed term agreement, you have to order more that you will use because you always want to have some reserve for some unexpected data. And the access to your data and for high resources images when you really, it is very hard to state before how much data file, how much images you will need before getting them. For example, you have some criterias on the data, the time, the satellite, obviously the area, but before you start getting the data, you don't know how much of the products are really covered by cloud or in other ways inaccessible for you. For example, they weren't created by the source, so it is better to pay for those projects after you have done your calculations. So, choosing the best option is somehow, you have to consider the base, some base, free base things. First, you need to ask yourself if you need the fixed term or if you can remain in paper use mode. Second is how long is your contract? Because short contracts, for example, one and a half months, the fixed term prices are per month, so you have to order two-month contracts to get the environment started and running, so it will be cheaper to use to run it in paper use mode and just shut it down on the end. And the scale of the contract, I mean the time frame of the contract is really the most important. If you have a short something quick, for example, when you are thinking in days, then the paper use mode and pay-as-you-goat modes are more user-friendly and they just, at the end most of the time, you will get the calculations cheaper than using the fixed term contract. And one special thing about the CRE-ADS is that we provide our users with the possibility to sign a six-month contract, which is something that most of our competition doesn't have. And the pricing is almost as good as for the one-year contract, so you can adjust it accordingly. And as I said before, there are three special kind of products, the storage, the product generation, accessing the data that mostly require paper use, but this is one thing that you should be aware that object storage should be really only used in paper use mode because in object storage, the special kind of storage that you can only, that you only pay for the data that is really there. So when you are ordering a volume storage, you must order the free space on the volume also. And realistically speaking, there won't be a situation when you are using 100% of your volume because then adding more data is impossible and you return machines or anything that you do, you crash the server and it stops working. So you have to always buy more. And in object storage, you don't really think about the free space, you only put more data, so the paper use mode is always better here. And when you are ordering from our website, there is a very quick and easy way to order the fixed term contract on the e-commerce, just there are four easy steps, you add the environment, you configure it and you are ready to start. And the longest term, the longest thing that is always there is paying for the invoice because in some organization it takes not seconds like every other span, but minutes and hours. And in all billing modes, whenever you are using the computing environment, you have always the access to the dashboard and to the APIs where you not always can see how big is your environment and what are you doing then and how much instances, CPUs, RAM, etc. you are currently using. But you can also start and just click and use more computing power. And also you have all the APIs listed there. So you can, this is basic open stack. So you can just do whatever you want there and script the whole process. And most of the big clients use Terraform scripts or other tools to really quickly start hundreds or even thousands of VM and later destroy them. And this is really the right way to use a cloud, not from the fixed term perspective, but from the API perspective and not managing it from user console and some graphical interface by APIs and by issuing computer understand comments to the cloud. Okay, so what is the right way to how to determine which one is the right model. So when you look at our price list for virtual machines, you can instantly see the price for the yearly contract monthly price is way lower than the price for monthly contract. And if you calculate the price per hour, you will also see it is higher than the per month contract. So the VMs have the resources that the VM gives is the same depending on the not the same, but you get the resources that are stated here. But you have to choose the environment and in the pay per use mode and pay as you go mode, you can always change the virtual machines just just skew one machine get another and that's that you don't have to write any new contracts. So this is the thing that you have to understand about the cloud environments that if you are in the pay per use mode or pay as you go mode, you just can if you come to understanding that this VR virtual machine is too small or too large, you can always kill it and get the other flavor to run and also to experiment and to develop new capabilities and to test them. But the prices as I stated for the storage are mostly the same regardless of how much you are using, how long or how long you'll be keeping the data. So the price initiative for the storage and order ranking for one year is virtually known, it's only a few percent. So most of the users use the storage in pay per use mode and the virtual machines in a mixed fix mode pay per use mode depending on the on the computation and depending on what you want to do. And also for some users because those previously offered virtual machines for some users having the whole server, compute server is very important. And for those users, we have a special line for virtual machines. Those are really dedicated servers with virtual machines. So each VM takes the whole compute server and you have the resources of the whole compute server just for you. It's not only the processor and disks, but also the network and network access to the AO data. And all of these machines comes with the local SSD drives. And those local SSD drives and NVMe drives are very fast and have very low latency. So you can expect up to three 100,000 IOPS from the driver even more depending on the flavor. And in some machines, we offer the GPUs. So if you are interested in GPUs, we also can provide those and we'll be expanding those often in some future. We are just generating information what is really needed. And we have also for some time the new line of virtual machines, the virtual machines with local storage. And they have the same price as the high memory line. The storage for those machines is inside the compute servers. And they have very low latency and very high IOPS. So if you are, for example, searching for the virtual machines for processing, it will be easier to run it on those machines and not on the machines that have a network storage. Because if you have a network storage, for once the storage consumes some of the network and for other, you have a bigger latency and latency is always back. And as I said, the price is in line with the normal high memory. So we basically can choose if you want to have this VM on a network storage or non-local storage. So just to a quick wrap of the things, what to consider when you are launching on the project on CREODS. The first thing and the most important thing is to the length of the projects. The second thing we encourage all of our clients is to test our environment, test your application to know what CPU power and other things you need. If you want to run a longer project, it is better to talk with ourselves guys, to work with the agreement that works for you and for us. So you can have a guarantee of your environment and sometimes a better price for computational resources. And it's always better to order storage in a paper use mode and because then you don't pay for all for what you are using and not for some free space or what you expected to use. And what to remember about what I said earlier, that each and every product can be built in different ways in fixed time only in paper use and you have to ask yourself what is the best option here. That you can work with our business developers and with ourselves guys and to work the best agreement that will help you and that will work for you better. We provide custom solutions and we improve the platform on a daily basis. For example, we add more data, we add more processors and if you are looking for more data and we can build some business case around adding more data, we can add it to the AO data. So if you want some data collection that is not here, start talking with us and we can work from there and try to add it. And we can also put your custom processor, your custom data on the Deus for everyone to use and there are ways to share also your data that you have in your private repository and create yes with other users. So you can collaborate with others and not have to copy the data every time someone wants to use it. There are many different ways to use cloud and cloud services and it is important to remember that they will change in time and that one model when you order the resources for a longer time is changing and in the future we see that the cloud will be going in the direction that the users will pay for the int data and the result not for just doing things. The CREO-DS is comparing to the market environment very flexible and very agile and we can do things really really fast if it is required and start new services in a matter of days and weeks, not months and years and we help our users and our customers grow. We help to share the data, we help to share the processors, share the applications, promote them and to show you how some use cases and how to get to the users and who can be interested or work together together towards the tender to start working on something. Thank you for attending and I hope you have a great day today. We had a solar eclipse today so this is one of those unique days. Thank you very much and see you on the next webinar.