 Good morning everybody. I'll just attempt to get my screen share all set up and then we will head off. So I'm hoping that you can see my first slide that says Jasmine overview in the middle. Is that correct? Fabulous. So, yeah, first of all, I'm Alex Stevens. So I work with Poppy at the Centre for Environmental Data Analysis. I'm the head of partnerships there. And my role involves various aspects of what we do, but significantly includes working with teams that want to access Jasmine and do their science on it. So I'm just going to check the chats to make sure there was nothing coming in from me. So Poppy, just to say, please interrupt at any point if people can't hear me or anything's going on like that. Keep everyone on board. Yes, so I just want to say a big thank you to all my colleagues at CEDA, but also at the Scientific Computing Department at STFC who help us co-manage and run Jasmine. So I'm going to give an overview of what Jasmine is, but I'll before then just start and tell you a little bit about CEDA about who we are. I'm going to talk about the Jasmine services that are provided and different types of usage pattern. And then I'll go into more of a typical individual scientific user, how you might get access to Jasmine, log into the system, work on the find software and work in our interactive computing environments, those kind of things. So briefly about how we manage and allocate resources on Jasmine because unfortunately it's a finite system, so we have to share it out fairly. And then I'll also mention a little bit about what we do in terms of support and outreach and feel free at any point to speak up and ask questions if they're relevant along the way. So we'll need to be interrupted as you may be asking a question that someone else wants to ask as well. Okay, so we'll just start by introducing you to CEDA. So, as I say, we're the Centre for Environmental Data Analysis. We are based at the Rutherford Appleton Laboratory near Digcop. Our mission is to provide data and information services for environmental science. And we have been going now for about 25 years. And we have around 30 members of staff, a mixture of data scientists and software engineers. And we are largely funded by NERC, although we do get funding coming in from a variety of sources. And there's just a nice picture of our 25th birthday party there. So going into a bit more detail of who we are at CEDA. So, as I say, we're part of the Science and Technology Facilities Council within UKRI. We hold a large petascale archive of both atmospheric and Earth observation data. And this comes from a range of NERC sources, but also many third party data sets of interest that we know the research community need to access to do their science. And our entire archive and data services are all housed on the Jasmine platform. And so in the context of this lecture series, we are part of the Environmental Data Service or EDS. And so the EDS is aiming to manage data created from NERC funded science or needed to help environmental research happen. And this is an amalgamation of a number of domain repositories. So here on the right hand side, we're the fourth on the list and CEDA, but I'm sure you recognize some of the other major players for different scientific domains in the UK. So the idea is that we are becoming more of a joined up service, which takes its time, but it's something that we're all working towards. Essentially, we at CEDA run two key services. So we manage the CEDA archive. So this is a multi petabyte archive of environmental data and a number of services and resources on top of that. And that works in parallel with our Jasmine data and compute cluster. And it's important for you to know that everything that CEDA runs is all housed on the Jasmine platform. And here are just some examples of ways in which you might access CEDA. So we have a catalog, so web based view of all our data holding so you can come in and search for for climate simulations, for example, or you can see the great scans might be interested in. And then you can drill down and find data either through our web interface, you might log in via an SSH terminal Jasmine and run your code against the data. And then we also have a Jupyter notebook service which I'll tell you a bit more about. We have other web tools such as a web processing service that allows you to select subsets of data that you're interested in, and then just pull down the data that you want. Okay, so that moves us on to Jasmine itself. So Jasmine is a compute platform that's operated by CEDA and the scientific computing department at SDFC, and primarily operated on behalf of NERC. It provides flexible storage and data environment that supports the need of a large range of workflows. Here we have some numbers in the bottom right corner. So we have over 14,000 compute cores, and we have I think in different parts of this talk I probably go beyond 45 petabytes, so order about 45 petabytes of storage the moment. And the Jasmine philosophy can be very much thought of as bringing the computation to the data. So in the olden days when we had small amounts of environmental data, it might be easiest to pull the data to where we are. In the olden days there is just so much data out there. The idea is that we essentially have a large data pool on Jasmine. And we provide the compute resources so that people can come on to Jasmine to do their work there. So I'm going to talk now for a little bit about an overview of our services and some examples of how different users in different projects might use those services. So who uses Jasmine? So at the moment we have at least 2000 users. We have over 300 different scientific projects which are using Jasmine for either part or their entire operation. There are a range of scientific disciplines as indicated by the image on the right here. We're primarily supporting UK usage but of course a lot of science is international these days. So we have a lot of European projects and globally international projects that use us. And we're primarily but not entirely about supporting academic research. So the idea of visualizing a sort of typical use of Jasmine is to think about the data analysis and sharing model. And in this graphic, we sort of talk through the different stages. So if we think first about the curated data archive at CEDA. We have a lot of data sets that our scientific team have identified as being important or the community has told us are important. And we go through a proper process of bringing those in creating catalog records, looking at data formats, etc. and curating the data. In place, a number of Jasmine users are able to log into the system and run their analysis on the data, and there are different ways they can do that and I'll talk about those later. That quite often leads to for certain projects and certain teams, the ability to actually process data sets in order to produce new types of data. So you might be bringing some of your own data to the system. You might also be analyzing or using as inputs data from the CEDA archive. And then typically those projects will used a managed cache and storage system these things called group workspaces. So for particular communities and projects we set up group workspaces and those can be many hundreds of terabytes in size in some cases. And so typically these are places where process data or new products would be created. There are different ways that those can be disseminated from there. But there's also a large a last feedback that happens in various cases where actually the outputs from this entire process are new data products and those data products are things that are important to the community that therefore might be pushed back into the curated CEDA archive for wider use and accessibility. So one of the things about Jasmine is always been quite a challenge to try and represent everything that we provide and everything we do. So this is one way of breaking things down and thinking about the issues and these these boxes and the color coding on this particular slide will be represented in the following slides. So hopefully you'll be able to follow that having seen a few of them. In the column here, we have a number of information services. So these are different ways that we can tell people about our resources tell people about what's available and allow people to discover how to use and work with Jasmine themselves through online resources. We have a number of access services. So there's an accounts portal and different ways of connecting to the two Jasmine, which I'll come and tell you about more in a few slides. There's a lot of analysis compute capability. So there are interactive servers, Jupyter notebooks, a batch cluster. And again, I'll talk more about these. Everything in blue is our storage and there are a number of different types of storage, different types of media with different usage patterns. We have something called our own community cloud. So we run our own cloud services, and we have a portal that allows projects to log in, have access to their own resources and build their own applications and systems within the Jasmine ecosystem. And then a significant amount of networking which is really underpinning Jasmine, connecting all these things together and allowing people to do sort of high bandwidth and big data science. So that's looking at the services by type. If we think of how they are arranged, and we still have our information services on the left hand side here. And then we've sort of broken things out down into different areas. So over on the right hand side, a significant amount of what we have here is using our cloud services. The top left here is about access systems, login and basic transfer. Everything in orange is our compute services. So this is this is where people are either logging in and running their analysis and their processing. Or scheduling jobs on our batch, batch cluster, which is called Lotus and running large amounts and large scale of potentially parallel jobs. So down the bottom here, this is all about storage, and I'll tell you a bit more in the next examples. So the first example, we can think of somebody bringing their scientific workflow to Jasmine. So you might be working your local institute or you might be working on your local laptop or something, and you're looking to scale up or access data on Jasmine. So part of the process is going to our accounts portal and signing up for access, getting an accounts portal account, which gives you the ability to request resources. You then ask for a Jasmine login account, which would then give you the ability to SSH into the system. And typically people start by logging into the size servers. These are just interactive terminal servers that you'd SSH into. And from there you can access a set of common software, which I'll tell you about. And also, then you can connect to the batch compute system. So once people have developed a workflow on the site, the interactive servers, they typically then move to the batch compute nodes. So they can run larger parallel batch systems. In order to do all of this, the user would have access their home directory, which uses solid state storage. So it's nice and fast storage and each user has 100 gigabytes of home directory available to them. But as they scale up their work and they want to work significantly larger scale, they're likely to need to use a group workspace or a GWS. And this is on a different, typically on a different type of storage system, SOF stands for scale out file system. And then also some users will want to write temporary data to our scratch file system. This is another large part of the storage system on Jasmine, which people can use for intermediate products when they're running their big data workflows. So all of this is showing how a user would come in and develop something and then build it up to run and parallelize on the batch system. Our second example, this is about transferring data. Typically, once a user's got a Jasmine account, if they if they need to bring some data in from outside or they want to transfer data from Jasmine to a remote location. They would use the transfer or X for servers, which are available to all Jasmine users. But in some cases, a project needs to do a large amount of high performance transfer. In that case, they might sign up for access to our HP X for system. So this is a special data transfer zone that we've set up. And it enables higher bandwidth transfers for those looking to transfer terabytes of data. And they can also access tools such as Globus online using that. And then typically they might want to use the XFC system, which is the transfer cash to bring data in. And then copy the relevant data to a group workspace. And then after that, they might want to put some data on the tape. If they've got got too much data that doesn't all fit on disk. And we have a storage migration service this thing called the JDMA that allows you to move data from to and from tape and disk. And then finally, once the data has all been brought to Jasmine, it's likely that there will be a requirement to use the compute service to to run their code on the batch system in order to process all this data. So the third example is about actually building applications on Jasmine and I have a few slides in a minute that show what that looks like. So a community might decide that they want to build some web interface that talks to data. And they might want to actually manage it in a way that they manage their own users and their own software systems so they can request access to this to an external tendency on our cloud system. And within that then they could build their own virtual machines or they could use this thing called cluster as a service. And that's a set of building blocks that allows you to to build applications on the cloud environment. For example, you can request a Pangeo service, and that will allow you to build a set of Jupyter notebook services with a load of software already installed that your users can then log in and use. And so users would log into the cloud portal in order to interact these things and select them. And then they're just taking you back slightly. And then in order to access other parts of Jasmine, they might be wanting to access our high performance object store this thing down here, which is a special type of storage, which has an HTTP interface. So instead of being logged into the location where the storage lives you'll be able to access it remotely. So let's just look at a few examples of projects that have used some of these things. So Primavera was a European horizon 2020 project that was about managing multiple petabytes of climate simulations. A group of over 100 scientists were using Jasmine as their main analysis and storage location for the project. Overall, they had to manage about two petabytes of data. And so that was organized on tape and disk. And at any point in time they had about 400 terabytes of data on disk. And the actual processing was all done on the Lotus cluster. And they had their own bespoke web application running on a virtual machine on Jasmine. And so this is all about understanding how really high resolution climate modeling allows you to resolve and understand process processes that aren't necessarily fully resolved in lower resolution models. Another project is run by the Center for Ecology and Hydrology or CEH is the NERC data labs. And this is using our cloud based system on Jasmine so it uses our infrastructure as a service and allowed this project to build its own set of components and to interact with each other. So they have a web browser that provides a sort of set up and user management system that can then access data in our object store. Users can bring their own code to it, create their own visualizations and then interact with the system using Jupiter notebook service. There's just a screenshot here of the types of things that users can do with the data labs. And this project is worth just talking about because it's a really good example of how everything connects together. So the DEFRA JNCC data cube project combined a number of services. So they start with Sentinel satellite data, which is being brought to Cedar on a daily basis terabytes of that coming in each day into the Cedar archive. So within the project they use the batch compute cluster in order to generate their own version of the data, their own analyzed data. First of all, written to the storage system, the group workspace, but then also copied into the object store. And this essentially allowed them to transform the original data into what we would call analysis ready products. So analysis ready data. So that's putting it in a form that makes it really easy to use for external users. The external users then use this data cube app, which is running inside the cloud tendency on the Jasmine cloud. And this uses the cluster as a service approach, which deploys a ready made Jupiter hub service so that they can then open that up to users, they can come in and run Jupyter notebooks and access all this Sentinel data. And applications for that are things like habitat detection and illegal waste detection. So I mentioned various times there are Jasmine cloud services is just a couple of screenshots of what you might see if your project was given a tendency on the system. So on the left hand side you're, you're looking at resources such as number of machines. The number of IP addresses available CPU, RAM, storage, etc. And on the right hand side, you're looking at a point and click interface to actually select pre configured building blocks such as Pangeo. Slurm is a scheduler and then different file systems that you might want to use. These are just examples of some of the tools that people we use on Jasmine. So I'm going to move now to a view more of how a user, an individual scientist might typically use Jasmine. So, first of all, they need to get access to the system. I'm going to point you to the help pages, and they give a step by step guide to how you need to get onto the system so that would involve generating an SSH key, getting access to the Jasmine accounts portal, checking your network details. That's important because there are some networks that may not be able to connect directly to Jasmine. These big networks are able to connect fine, but it's just worth checking those things. And then the user gets an SSH login and can apply for additional services. So for example, I talked about these group workspaces, which are these large allocations of disk for a given project. So typically if someone starts on Jasmine, they might be connected to a project which has a group workspace and they would need to apply for access to that. So, your first login is probably via SSH and via an SSH terminal. So that would come from a client on a remote machine, it could be Windows, Mac, Linux, and each user has their own home directory which is mounted across the system. The login service allows them to then connect through to other parts of the system such as the interactive servers, the batch servers, and other bespoke servers. And this is just an example of what you might type as an individual user to log in to Jasmine via SSH. So we're setting up an SSH agent session and then logging into the login server and the login server will give you a list of the interactive scientific analysis servers that are available and tell you about the resource being used on each one. And then you can identify one and log in. For example, we're just setting up a software environment, and then running a Python session importing some Python packages that you might be familiar with. So moving on to software, the aim on Jasmine is to provide a common set of software for data analysis. There are a lot of help pages at the top level, we kind of break this down into different areas of interest. So there is the software, the open source data analysis software that's available across all the interactive and batch servers. But there's also tools that are there for compiling and building software, such as specific compilers, you might need for working on parallel code on Lotus. So that's restricted to certain users and certain machines. And then there is software that's specific to certain servers that's available to everyone, and data movement software which I'll mention as well. So the main software you'll be interested in is the software that's available across all the interactive servers and the batch cluster. And this is quite a lot of Python software but also a number of other packages. And these are packaged together into separate software environments. This tool called Condor as the package manager. So this allows us to have multiple software environments that you can switch on or off. And within them we've tried to pull together a set of very common open source packages people use for environmental data analysis. So things like CDO, XRA, Matplotlib, etc. And also some packages we install via RPMs, those that we can't install via Condor. And those are known as this Jasmine Psi environment, whereas the main environments are known as Jaspy and they're all documented on our help pages. In terms of other software, there are other languages available on the system if you need them. We can specifically pay for some IDL licenses because various people in our community use IDL both interactively and in batch mode via runtime licenses. We have some workflow management tools on a specific server that allows them to connect to Lotus and manage batch processing. These tools called Rose and Silk that are developed and provided by the Met Office. And we also have some community contributed tools such as ESMVAL tool and Afterburner. In terms of data transfer, there are standard command line tools available such as scp and rsync that allow you to copy data to and from Jasmine. And I mentioned that we have this dedicated high performance transfer server as well. If you're using that you can use grid FTP or Globus online for getting greater bandwidth. And one thing we always want to ask is do you need to copy data. So there are, on Jasmine, a lot of the storage is mounted across all systems. So if you're working in one part of Jasmine, you might actually have access to the same data on another part of Jasmine, which means that you don't need to copy data from one group workspace to another, for example. And certainly if data is in the cedar archive, you can read it directly and you don't need to copy it to anywhere else. I've mentioned a number of times this batch processing system. And so this is this is called Lotus. And typically, if your workflow is long running or complex, you will want to run it on Lotus, which is our batch compute cluster. So we have over 400 compute nodes, around 13,000 compute cores. And schedule a tool called SLAM, which manages all the jobs and tries to manage the load fairly across the system. So there are multiple queues for different types of jobs, such as running very high memory jobs or multi processor jobs. And we would encourage all Jasmine users to move towards using Lotus for their processing their analysis. In terms of the amount of usage of Lotus. So here on this graph, we're just showing millions of CPU hours. And this, this is just showing how we transitioned from our previous scheduler, which was this tool LSF to the new tool SLAM. But you can see that on average, we've probably got about two and a half million CPU hours per month running on our batch compute systems. Another service that I mentioned was our Jasmine notebook service. So for those of you that aren't familiar with them, the Jupyter notebook is an interactive programming environment that runs inside a web browser. And within a notebook, you can define, edit and run code. So we, we run a service in Python, there are some that also run in R and other languages. And you can embed visualizations within your code and documentation. And you can very easily share your notebooks with others via tools such as GitHub. So they're really good collaborative tools. And hence it's really useful to set this up on Jasmine. We will, we'll share these slides with everyone. And there is just a two or three minute video that just shows going through the process of writing some Python into a notebook environment. Importing packages, reading data in the Cedar archive, and then generating a visualization interactively from there. So overall, the Jasmine notebook service is a it's available to all Jasmine users and be it's providing an alternative to an SSH terminal access. So it provides a Python programming environment. It provides all the same data analysis software that I said was available on the interactive servers and on the Lotus cluster. It's also connected directly to the Jasmine file system. So you can read content from group workspaces and you can read directly from the Cedar archive. Okay, so I'm going to briefly mention how we manage and allocate resources. And as I say, unfortunately, Jasmine is a finite resource. So, from the perspective of your project, you typically come in related to one or more projects, and there will be a project PI or lead who is managing your project and and your interaction with Jasmine. So that project will get in contact or the PI will get in contact with one of the Jasmine Jasmine consortia. So there are a number of consortium managers who have an understanding of various domains. So atmospheric and polar science, for example. And those domains have an allocation of resource available on Jasmine. So this will be an amount of storage and compute resource object store that can be made available for different projects within that particular domain. So the project PR will talk to the consortium manager. And through that process will put in a request for some allocation of resources. So in terms of group workspace, they might ask for different access to different media, such as disk tape and object store. And there'll be a responsible party within the project called the group workspace manager, who will be in charge of organizing use of that, that particular allocation of resource within the project itself. And then again, the project might also be interested in in having a cloud tendency, in which case there will be an administrator within the project who liaises with with the Jasmine team. And that cloud tendency will have an allocation of a number of cores, for example, a certain amount of local disk allocation, potentially object store allocation as well. So I think you've got about five, 10 minutes left before. Okay, questions. All right. Thank you, Poppy. And there is a Jasmine projects website that people who are accessing Jasmine resources get an account on. And through that you can find out about the consortia set up a description of a project and request different resources through that. So I want to also touch a bit on support and outreach, so how we communicate with the world and the kind of tools and support that we're providing. So we have a significant amount of support just to help.jasmine.ac.uk. So this is where we try and explain what all our services are and signpost people for how you find out more information and how you can sign up for access, etc. And then we also have on the CD site a news blog. So this is where we try and keep people informed of changes and also point to updates and things such as events that that might be interesting to Jasmine users. So we also run our own training. We run Jasmine workshops. Initially we were running these in person. In the days where you used to actually have to stand up when you were presenting. That was always really nice to sit in a room and find out from people what their requirements were what they wanted to do with Jasmine. So we've been able to migrate that into an online version, which we, we run periodically but we've also put all the main resources on GitHub and YouTube. So for many people if they want to go and find out they can follow the links and actually watch some of the videos. So the idea is that we've split this up into a set of exercises. So that we, for example, just connecting to Jasmine as one of them, and there's building your own Python three environments. And so the idea is we present an exercise, and then you can go away and actually attempt to do it yourself. And then we explain ways in which you may have done it. So we run a set of webinars. So on the CD site again, there's a list of previous webinars and new webinars are advertised. So from time to time we'll run webinars in things like getting started on Jasmine, or working with the batch processing system on Jasmine. So we run an introduction to scientific computing course, which is primarily about getting to grips with Linux, the basics of Python, and then starting to talk to people about using common Python packages to support scientific data analysis, particularly in earth and environmental sciences. So in other ways, we try and support and connect with the community. So we run a series of Jasmine user seminars. And this is a great way to share information between projects. So I said there are 300 projects on Jasmine. This is an opportunity for people from different projects to actually come and tell us about what they've been doing and share with the wider community of Jasmine users, how they're using Jasmine. So that's an ongoing series. We also in 2021 supported a number of hackathons. And so we had people and communities who are preparing some data analysis for COP 26 meeting who came to us and said that they would like to use Jasmine as the location for running their shared code and their shared data and using the Jupyter Notebook services. And so we were able to set up temporary resources to support those particular activities. Last thing I want to say was that at CEDA we are leading a project for UKRI, which is about targeting net zero for the UKRI digital research infrastructure. And this is really interesting from a Jasmine point of view because it's significant part of it is trying to work out how we can make the most and best use of the compute resources we've got. So how can we reduce and limit storage? How can we make compute as efficient as possible? And some aspects of this are very much about good practice. They're about understanding the available resources and making sure that workflows are doing the appropriate things to minimize duplication and minimize having to rerun significant parallel workflows. There are a few links here just to follow up. We have the Jasmine site at the bottom is just jasmine.ac.uk and there are various help pages here. If you just want Jasmine help it's just help.jasmine.ac.uk and you can follow various links from there. If you don't find everything that you need, you can go and find us via our help desk. So we are support at jasmine.ac.uk. If there's a CEDA question you want to ask it's just support at cedar.ac.uk and you can also follow us on Twitter or find out more via our website. So thank you very much and I'll pass back to you Poppy for questions. Brilliant thanks so much Ag that was great. You always give such a clear overview talks about Jasmine so thank you so much for that. So just a reminder for everyone that there's a Q&A box that you can add questions to. We don't have any at the moment but I'm sure you're probably all just thinking it through because there's a lot of information and we know Jasmine's quite complex. So I'll just keep an eye on that. Okay so we do have a question. So we've got a question which says how does EDC I guess for that you mean the the EDS which is the Environmental Data Service that's the wider collection relate to CEHEIDC and is there information on costs that we might need to put into grant proposals or is it for free. So I also should have said we've got Sam Papla here who is a curation manager at Cedar so he's going to jump in if we need any help with questions and I think Sam that sounds like a you question so do you want to give a bit of an overview about how the different data centers link together to make up to the. Yeah, so yeah. EIDC is part of the EDS the Environmental Data Service in exactly the same way as the Cedar archive is. The Cedar archive sits right on top of Jasmine but the EDS doesn't so they're still, well, not entirely anyway because there are still pieces of infrastructure in CEH but it all works. We all work together. The question about costs and proposals I think it does it does often depend on for all the EDS data centers when if it's your front of the cost of archiving some data, then most of the time it's basically free at the point of delivery so but in the exceptional time it's enormous, or has some other special requirements we did those costs are often factored in, but currently they most of the time, if it's, if it's reasonably small, where reasonably small varies, that's one of the troubles we have. Then it's just, it's generally okay I think one of the key things we have put into the grant system is there is an online data management plan where at least you get a list of data sets so that we know what's coming so we can answer those questions like proactively rather than actually wait until somebody tries to give us something. It's just enormous with with various other problems associated with it. So that's generally how it works. And hopefully that answers that one. If not, feel free to send in any extra question to clarify. And so I think we've got another question which says, what minimum level of skill do users need to be able to use Jasmine. That's a very good question, Ron. I, I think one of the, one of the things we realized early on is that some people are coming to Jasmine, having never used Linux having never used an SSH terminal. That was part of the reason why we, we created the Jasmine workshop approach, because it comes in at a very basic level and says if you, if you've got no idea we'll talk you through the steps. But, but the bit using Jasmine means many different things to different people so some, some people will be users of services on Jasmine that mean they never even have to get into an SSH terminal if they could write a bit of Python, they could, they could use the Jupyter network service only. And some users will use a service like data labs, for example, and not access any of the core Jasmine services. If we're specifically thinking about people who are going to be running reasonably large parallel workflows. We always like them to have a bit of a bit of experience and a bit of training before we let them loose on being able to maybe sometimes in use resources inappropriately as it is a way I'd put it politely that unless people have got experience and guidance on how to use the resources, they may be doing things that are very, very inefficient. And just keeping an eye on the chat. I'm going to put you on the spot here a little bit, Sam. But, so I just wondered whether you wanted to kind of explain how Jasmine and the EDS kind of relates because it's not it's a bit unclear sometimes isn't it how the, the data centers link to Jasmine so maybe try and summarize that to keep it clear in everyone's heads. Yeah, so I think it's a Jasmine is definitely the computer system and the associated services about using the computer system and its storage. And, and currently the, the EDS is is the seed of it is entirely set upon that. We use that storage but you know in history before Jasmine we still have data so we have moved on to it. Perhaps there is a future where Jasmine doesn't exist and we will have moved off and maybe set signal data somewhere else and using different resources. So, but currently, Jasmine is the best game in town, certainly for the Cedar archive, and, and NERC is committed to having it run is in its, in its strategy, Jasmine is written through a lot of the kind of digital information that I think quite heavily. There's definitely going to be support for Jasmine for a number of years yet. So, I don't anticipate moving off it. A lot of the other pieces of the data. The environmental data service are thinking about moving on to it more so so the data labs thing that we flag there is clearly something that the people at CH generated and is there, and they're moving more of their infrastructure over. But we've also talked to people at Bass about moving things, including their kind of processing over but you know do so people are thinking about how to how to move things over to it because they can see it's a supported system, which is going to have a long enough. It's gone a long enough future in front of it to actually go okay this is how we're going to do it. So I think it's the way NERC has laid out okay this is how we're going to support that infrastructure. And I think that they're still it won't, Jasmine doesn't do everything you need. So there are other components that we need and I'm anticipating those existing as well for a long time, but it's certainly a good place for doing the things it does well. Large chunks of storage. It's clearly a cloud kind of infrastructure which we're all getting used to more, and it's clearly got a big batch compute so if it's those three things, then I think lots of things that the EDS do is going to migrate that way. Thanks. That's helpful. And I spotted another question coming. Let me just find it. So it says, what kind of support is there to help make running large jobs, jobs more efficient. So I guess this is about large scale data processing. Yes, I've got a screen share. Thanks, I've just just thought it would be worth pulling this up. So thank you for the question, Michael. So this is our GitHub resources for the Jasmine workshop. And part of what we set out to do when we built the Jasmine workshop was to push the idea of good practice and using resources appropriately. If we start from the very beginning, which is connecting to Jasmine, using the side servers transferring data to them from Jasmine. We then we kind of take users through the approach of you learn how to do something interactively and then you put it on the batch compute. And as part of that, some of these exercises talk about where you might write your outputs. And then we talk about things like building your own pipe, but your own software environment. We discuss the most appropriate place to do that for efficiency and things like that. And we also, we haven't migrated this one online yet, but this is an exercise which is about. Describe a workflow situation that involves processing and then reprocessing and 20 terabytes of data something along those lines. And then it's a kind of discussion exercise about the most appropriate way to use the different resources. So, so the Jasmine workshop is one area where we've tried to document some of these things. In reality, every workflow is different. So we would encourage everyone to talk to our help desk or talk to people in the team that you know. In order to get advice on, you know, the best the best way to break down your big task into lots of little tasks. The best way to use different parts of the storage because some of them are much more efficient for say large files. Some of them are much more efficient for lots of small files and software environments and those kind of things. So it's definitely worth if you're coming coming to Jasmine with a significant bit of processing that you need to do. So I think engaging with us and we'll do our best to to provide some pointers on on the most efficient use of resources. And I think that's something that kind of links to the net zero project as well as an ag where we'll be talking about. Yeah, what do we need to what do what the researchers needs to be able to make their workflows more efficient and that will kind of be coming into the evidence that we collect from that and proposals for what we think you can I needs to do. Yeah, I think if if you ask me what the solution was now Michael I would say, give it give us another 20 people in the Jasmine team, whose job is literally to sit down with whichever group is developing a new workflow and analyze it all. Look, look for bottlenecks look for duplication. And, you know, it, there's I wish that we had a lot more resource to actually engage really closely with all the projects and help them on this issue, but we don't yet. Thank you. Can you just check the chat, but I haven't seen anything else. So I think, unless anything comes in in the last couple of minutes we should probably say that's the end of the webinar so thank you very much everyone for joining. And just a reminder that there's another two webinars in the series says one on the fourth of February about the introduction to the next vocab server there's lots of information on the website to go have a look there. There's another one after that on the 26th of February, which is about data lab so that's the tool that I mentioned earlier so if you want to find out more about that then register for that one too. And just want to say thank you everyone for coming and listening to us today and hope you have a good day. Thank you everyone thanks Bobby. Thank you. Okay.