[00:00.000 --> 00:17.880] Okay, we're ready for the next talk by Courtney who's going to talk about open source radio [00:17.880 --> 00:20.360] astronomy. [00:20.360 --> 00:21.360] So hi everyone. [00:21.360 --> 00:27.440] I work for Ostron which is the Dutch Institute for Radio Astronomy, we're a government institute [00:27.440 --> 00:30.600] so we're publicly funded. [00:30.600 --> 00:34.560] Apart from being publicly funded we have a lot of research grants and that's basically [00:34.560 --> 00:40.040] what pays my salary and I'm going to talk about our biggest instrument that we have [00:40.040 --> 00:45.480] which is called LOFAR and we actually utilize a lot of open source software there, almost [00:45.480 --> 00:48.000] exclusively there's some caveats. [00:48.000 --> 00:52.320] My name is Courtney Leuker by the way, I'm also into amateur radio, my call sign is [00:52.320 --> 00:57.240] Papa Delta 3 Sierra Uniform, this is going to be quite a relaxed talk, we're going to [00:57.240 --> 01:01.200] give a high level overview of all the different components that are there, there's quite [01:01.200 --> 01:06.200] a lot of them actually so it's not possible to go into detail with the time we have. [01:06.200 --> 01:11.560] This is my first forced them talk ever and also my first talk ever in a lecture hall [01:11.560 --> 01:14.800] so that's quite interesting. [01:14.800 --> 01:19.240] Now some background, I mentioned that we're a government institute, we firmly believe [01:19.240 --> 01:24.480] public money means public code and we stand by that in almost everything we do already, [01:24.480 --> 01:29.120] we have an open source committee that also ensures that we do that and we have basically [01:29.120 --> 01:34.520] two very big telescopes, one's called LOFAR which stands for Low Frequency Array and the [01:34.520 --> 01:39.400] other is the Westerbork Synthesis Telescope also called WSRT. [01:39.400 --> 01:45.120] There's some sister institutes that you work closely with or are related to us, one is [01:45.120 --> 01:50.480] called Kamras which maintains a telescope that we've stopped using and the others are [01:50.480 --> 01:54.720] Jive and EVN, I'm not going to talk too much about those today. [01:54.720 --> 01:59.600] What I want to tell you is that there is this principle that our radio telescopes work on [01:59.600 --> 02:05.680] that is called very long baseline interferometry and this enables us to do radio astronomy [02:05.680 --> 02:11.520] in a way that wasn't possible with traditional radio telescopes. [02:11.520 --> 02:17.120] This is the whole map of LOFAR, there's 54 stations in total, roughly 25 of those are [02:17.120 --> 02:18.840] located in the Netherlands. [02:18.840 --> 02:24.280] I say it's around 2,000 kilometers in diameter but that's no longer true because the one [02:24.280 --> 02:29.960] in Rosen is new and we're now about 2,500 kilometers in diameter. [02:29.960 --> 02:35.440] This diameter is also called the baseline which is where the very long baseline interferometry [02:35.440 --> 02:36.440] comes from. [02:36.440 --> 02:42.880] If we then zoom into a station you see all these tiles and you see these little squares [02:42.880 --> 02:49.520] and those are the different types of antennae and that is what makes this type of radio [02:49.520 --> 02:54.680] astronomy so interesting is that we don't have a single antenna to catch radio waves, [02:54.680 --> 03:00.720] we have lots of them, about 20,000 in total actually which is quite substantial. [03:00.720 --> 03:07.200] This center is called the super-turb and it's located in Exlo, the Netherlands. [03:07.200 --> 03:09.080] How can we actually combine this data? [03:09.080 --> 03:15.120] I told you that traditional radio astronomy relies on a parabolic dish or a single antenna [03:15.120 --> 03:19.160] and we try to scale those up, make them bigger and bigger but of course physics are at the [03:19.160 --> 03:25.080] limits at some point, you can't make a structure from steel that's like 500 meters in diameter. [03:25.080 --> 03:33.040] What we do instead is we combine smaller antennas to act as if they are a parabolic antenna [03:33.040 --> 03:37.720] and the trick about a parabolic antenna is that all radio waves, no matter where they [03:37.720 --> 03:43.520] are incoming from, they all have an equal distance to the receiver so we need to emulate [03:43.520 --> 03:47.400] that with our antennas and we do that in two ways. [03:47.400 --> 03:53.400] That is an artificial delay, an analog artificial delay by just making the line that it needs [03:53.400 --> 03:59.360] to travel across on the PCB or the coax cable longer but we can also do it digitally after [03:59.360 --> 04:05.520] the data is being sampled and then we can aim into the sky and create a very narrow [04:05.520 --> 04:10.680] beam that observes a very small portion of the sky and that allows us to zoom really [04:10.680 --> 04:16.520] deep into space and make very detailed images. [04:16.520 --> 04:18.840] But what is this radio waves actually? [04:18.840 --> 04:19.840] What are those? [04:19.840 --> 04:20.840] What are we observing? [04:20.840 --> 04:25.720] There are two types of radio waves that are being emitted by objects in space and the [04:25.720 --> 04:26.720] galaxy. [04:26.720 --> 04:31.720] We're only going to describe one phenomena today that's called synchrotron radiation. [04:31.720 --> 04:36.680] Basically if you have an ion, a charged particle, you accelerate that, then it starts creating [04:36.680 --> 04:38.600] radio wave emissions. [04:38.600 --> 04:44.280] The frequency and the intensity at the frequency that is actually very dependent, that's all [04:44.280 --> 04:49.280] details that are not very interesting for this talk, but one of these entities that [04:49.280 --> 04:53.920] emit these types of charged particles are sometimes black holes and we'll see an example [04:53.920 --> 04:56.320] of that at the end. [04:56.320 --> 05:01.040] So I mentioned black holes, there's other types of radio astronomy that are very interesting. [05:01.040 --> 05:06.320] We can also model our own ionosphere or enlightening. [05:06.320 --> 05:10.920] Pulsars are pretty interesting, these are stars that are rotating at a periodic interval [05:10.920 --> 05:16.600] and they have very strong radio waves coming from the poles of those stars. [05:16.600 --> 05:18.320] So what does Lover actually look like? [05:18.320 --> 05:23.640] A very small antenna as I told you, we can see on the left that there's like wires attached [05:23.640 --> 05:28.720] to those poles, those are actually dipole antennas and if you configure them like this [05:28.720 --> 05:32.960] where they are like a V-shape, they are called inverted Vs. [05:32.960 --> 05:37.000] These are the low-band antennas and on the right side you see the high-band antennas, [05:37.000 --> 05:41.680] they're like a clover shape, like a tie shape. [05:41.680 --> 05:48.280] Then we combine all these antennas, low-band antennas and 69 high-band antennas in a station [05:48.280 --> 05:53.160] and we send the data at around 3 gigabits per second to our HPC clusters. [05:53.160 --> 05:58.400] There's a two-phase cluster here, the first is GPU processing where we do correlation [05:58.400 --> 06:06.440] and beamforming and the second is a central processing which is more like CPU based. [06:06.440 --> 06:12.520] In the early days our computing cluster looked something like this, we had IBM BlueJeans [06:12.520 --> 06:18.240] machines which were based on PowerPC and they had a 3D torus interconnect which is actually [06:18.240 --> 06:22.080] a quite interesting interconnect. [06:22.080 --> 06:27.160] This was problematic because utilizing the floating point vector extensions required [06:27.160 --> 06:31.840] manually rewriting assembly which wasn't that nice and it was pretty hard to find developers [06:31.840 --> 06:34.480] who were willing or capable to do that. [06:34.480 --> 06:45.800] So we moved to commodity hardware, GPUs, two CPUs per socket, two GPUs per server, two [06:45.800 --> 06:51.520] GPUs per server, lots of networking. [06:51.520 --> 06:57.520] That's really what you see here, we had 32 gigabytes of 10 gigabit Ethernet and then [06:57.520 --> 07:05.720] in 2018 when we upgraded we had 24 times or 23 times of 100 gigabits in FiniBand but you [07:05.720 --> 07:11.840] also see that there's a lot of 10 gigabit Ethernet per device and I'm gonna go into [07:11.840 --> 07:16.400] that wider this in a minute. [07:16.400 --> 07:22.320] If you look at the data flow or more like a software site then you see that the antennas [07:22.320 --> 07:29.000] have ADC so these conferred the analog waves that are incoming to digital signals and then [07:29.000 --> 07:36.400] we do beamforming on the stations and we send data to the correlator and this correlator [07:36.400 --> 07:44.880] also does the correlation afterwards and you can see that once the correlator is done [07:44.880 --> 07:50.200] with it we store this to disk and once it's stored on disk then it's made available to [07:50.200 --> 07:52.440] the central processing. [07:52.440 --> 07:58.360] So the correlator and our GPU cluster cobalt are doing like streaming and the central processing [07:58.360 --> 08:02.720] is more like your traditional HPC. [08:02.720 --> 08:07.120] When we look at the data flow in cobalt there's all this incoming 10 gigabit Ethernet and [08:07.120 --> 08:13.760] this is why we have four or three 10 gigabit Ethernet links per cobalt server. [08:13.760 --> 08:19.120] They are streaming the data and we configure per station where it needs to send its data [08:19.120 --> 08:20.960] to. [08:20.960 --> 08:28.400] Then once it's there it's being transposed at roughly 240 gigabits and once it's transposed [08:28.400 --> 08:33.440] we do have two pipelines that essentially run in parallel, one is correlation and one [08:33.440 --> 08:38.200] is additional beamforming so we actually beamform twice in a sense. [08:38.200 --> 08:42.960] It's little bit more complicated than I'm sketching here but I'm keeping things simple [08:42.960 --> 08:48.840] because stations also have the capability to not beamform and send unbeamformed data. [08:48.840 --> 08:52.840] We have a special buffer that's called the transient buffer where we dump raw samples [08:52.840 --> 09:00.120] and can send those two clusters but the general pipeline is what I'm sketching here. [09:00.120 --> 09:05.720] If I assume into these two pipelines the correlator pipeline and the beamformer pipeline I don't [09:05.720 --> 09:09.680] want you to look at the details too much here because that's not interesting and I really [09:09.680 --> 09:15.120] don't have time to explain it but the trick is almost everything you see here is based [09:15.120 --> 09:18.920] on signal processing, digital signal processing. [09:18.920 --> 09:19.920] That's what we're doing. [09:19.920 --> 09:25.080] We're using the fast Fourier transform, finite input response filters and transforming the [09:25.080 --> 09:30.160] data in like the frequency domain if you will. [09:30.160 --> 09:36.320] Then it's put back into CPU memory at cobalt, some final transformations are being placed [09:36.320 --> 09:41.320] and then it's put into the disk so that's how it can work on it. [09:41.320 --> 09:47.840] At Ostrone we do a lot with software and I've showed you now how the data flows but I haven't [09:47.840 --> 09:52.640] told you what software components are making that data flow happen. [09:52.640 --> 09:58.560] For cobalt, it's actually one solid product that lives in the low-fiber repository. [09:58.560 --> 10:04.920] Please don't all now visit or get lab instance at once because it will die if you do that. [10:04.920 --> 10:08.520] Try to spread that out a little bit over the day. [10:08.520 --> 10:16.160] I'm sure I will upload the slides soon after so you don't have to remember all this. [10:16.160 --> 10:20.720] Then this, all these tools that are listed here basically except for cobalt with lifts [10:20.720 --> 10:27.200] in the low-fiber repo, those are more like what you would find on the SAP side of things. [10:27.200 --> 10:32.400] I'm going to explain, I want to address that this is just the tip of the iceberg. [10:32.400 --> 10:37.800] On our GitHub repo you can find a lot of more stuff that you can play with and I would encourage [10:37.800 --> 10:40.480] you to do so because it's quite interesting. [10:40.480 --> 10:45.320] Then there's also CASA core which is heavily being developed at Ostrone as well but it's [10:45.320 --> 10:48.240] not actually just a product just by us. [10:48.240 --> 10:56.200] A competitor or like a friend of CASA core would be AstroPie, very widely known packages [10:56.200 --> 10:58.840] in the industry. [10:58.840 --> 11:03.280] If you look at the radio astronomy tool kit, so the necessary things to go from antenna [11:03.280 --> 11:06.920] to image if you will, then these are your friends. [11:06.920 --> 11:15.160] There's DP cubed, WaySW clean, and IDG, and Rattler, AO flagger and I'm not going to talk [11:15.160 --> 11:17.080] too much about every beam. [11:17.080 --> 11:18.200] What do these things do? [11:18.200 --> 11:20.360] What are we looking at here? [11:20.360 --> 11:25.080] DP cubed is where you define a complete pipeline, so you have the incoming data, you need to [11:25.080 --> 11:31.440] do some transformations on the data, maybe you want to identify some noise sources that [11:31.440 --> 11:36.480] might be in your data, and eventually you want to create an image, and for this imaging [11:36.480 --> 11:40.080] you need deconvolution as well, and you also need gridding and de-gridding. [11:40.080 --> 11:45.840] So, AO flagger is where you identify noise sources, this can be anything, like radio [11:45.840 --> 11:51.520] instruments are very sensitive, so one noise source in particular would be electric fences, [11:51.520 --> 11:58.080] windmills, solar farms, bed quality LED lighting. [11:58.080 --> 12:04.400] Then we move to the imaging part with WaySW clean because when you have a radio telescope [12:04.400 --> 12:10.640] consisting of many small antennas, in between your antennas there are holes, and that means [12:10.640 --> 12:16.000] that you're not receiving the data as if you would have a very large parabolic dish, there [12:16.000 --> 12:17.560] are some differences. [12:17.560 --> 12:22.080] This creates some kind of fringing in the image that you need to filter out, and that's [12:22.080 --> 12:28.400] what WaySW clean together with Rattler and IDG are doing. [12:28.400 --> 12:35.800] In IDG is your translation from the data domain to like a domain that is useful for radio [12:35.800 --> 12:39.000] astronomical imaging. [12:39.000 --> 12:44.160] So I talked a little bit about CASACOR and how it was industry widely used, it's based [12:44.160 --> 12:48.840] on all the packages that have been around for a very long time, but we've actually [12:48.840 --> 12:54.040] switched it around now, so now CASACOR is built on these older packages, on top of these [12:54.040 --> 12:59.520] older packages, rather than CASACOR depending on these older packages. [12:59.520 --> 13:04.560] There's several unique features here, the UV domain is pretty interesting, so that's [13:04.560 --> 13:13.880] the domain about having your, about the plane that is filled, so those holes in your surface [13:13.880 --> 13:16.440] area if you will. [13:16.440 --> 13:20.600] And there's some phytonda bindings here, so these are all very nice tools that you can [13:20.600 --> 13:23.840] just play with. [13:23.840 --> 13:28.320] We also use a lot of open source tools, and we're doing quite well, there's still some [13:28.320 --> 13:35.400] close source software, I'll get into that in a minute, so we use OpenMPI, OpenMP, Slurm, [13:35.400 --> 13:42.000] GitLab, Grafana, and actually the part that I work on is PyTango, which is a SCADA system. [13:42.000 --> 13:47.320] So with supervisory control and data acquisition, that's basically your interface that we have [13:47.320 --> 13:52.920] on the individual stations, and those stations then configure underlying hardware with the [13:52.920 --> 14:00.920] antennas and the ADCs, and they report how they are configured to a higher level system. [14:00.920 --> 14:07.640] We also use Prometheus, Docker, and a variety of interesting open source tools, this is [14:07.640 --> 14:12.040] just the tip of the iceberg as well, there's much more. [14:12.040 --> 14:16.480] Next to our SEP cluster is also pretty interesting, which is actually where we use Slurm, we also [14:16.480 --> 14:23.760] have a DustSys cluster, which is a cluster shared with many universities within the country. [14:23.760 --> 14:30.960] Things where we can improve, well, we use CUDA, so that's not really open source compared [14:30.960 --> 14:37.400] to OpenCL or Falcon, we're using WinCC for monitoring, maybe you've heard of that package, [14:37.400 --> 14:43.160] it's Windows-based, that's why it's called WinCC, we're trying to face it out for Grafana [14:43.160 --> 14:47.760] and Prometheus, that's going quite well, I'd say. [14:47.760 --> 14:53.200] We have a lot of closed source SEPDA vendor blocks, so if you have your silings or what [14:53.200 --> 14:59.520] have you or your Altera, then they for instance offer IP blocks to implement 100 gigabit ethernet [14:59.520 --> 15:06.760] interfaces, and they're not too keen on you sharing those with the whole wide world. [15:06.760 --> 15:10.480] Then InfinityBand firmware is pretty closed source, I believe there's open source versions [15:10.480 --> 15:15.880] of that, but I don't think I know if they work quite well, and then the main area that [15:15.880 --> 15:21.400] we're actually struggling is with office management tools. [15:21.400 --> 15:26.480] This is definitely the area that we can improve the most at Astrone, we use Office 365 Slack [15:26.480 --> 15:31.480] and Zoom, and as you can see, Copano, Metamode, Jitsi, there's definitely open source alternatives [15:31.480 --> 15:36.040] to this, so there's no real reason why we should be using this. [15:36.040 --> 15:40.560] Of course, you need to host the infrastructure, and that also costs money, so there's some [15:40.560 --> 15:43.880] little way there, I'm not saying that it's definitely cheaper, but there's open source [15:43.880 --> 15:46.880] alternatives to this. [15:46.880 --> 15:52.200] Now I want to show you, I told you about IDG, that does the gridding and de-gridding, I [15:52.200 --> 15:57.880] told you about WSWClean, and the Dravartler part that does the deconvolution now, and [15:57.880 --> 16:01.640] I want to show you how those tools work in practice. [16:01.640 --> 16:07.920] So we have an IDO point source, this is our most IDO radio source that can possibly exist, [16:07.920 --> 16:12.920] it creates a very sharp point in the sky, we put it toward the gridding, and we get [16:12.920 --> 16:14.480] a point function. [16:14.480 --> 16:18.800] What this basically is, is this is the error of our telescope, so we now know, okay, this [16:18.800 --> 16:24.520] is the error it's going to generate in our images, because we don't have complete filling [16:24.520 --> 16:30.720] of the UV plane, there are holes in between our antennae, and then we can use the WSWClean [16:30.720 --> 16:36.640] image together with Dravartler for deconvolution to create an iterative process in which we [16:36.640 --> 16:39.760] iteratively remove the noise from the image. [16:39.760 --> 16:46.360] So actually I'm going to see, oh yeah, that's nice, so here you see these lines, these lines [16:46.360 --> 16:52.240] are the fringes that I've told, and if you then perform these iterative cleaning process [16:52.240 --> 16:57.800] on what are called calibrated visibilities, then we iteratively see that this image is [16:57.800 --> 17:01.280] drastically improved. [17:01.280 --> 17:06.360] So now some example of this, what is the science that we do with Lofar, how does this look [17:06.360 --> 17:07.360] like? [17:07.360 --> 17:13.200] Well this is the, this is a paper by Erk Timmermann, so you can look at it when you spare time, [17:13.200 --> 17:18.240] and what we're basically seeing here is we're seeing huge jets of this synchrotron radiation [17:18.240 --> 17:22.880] emissions that are talked about, and you can see that they actually over millions of years [17:22.880 --> 17:28.040] they vary in intensity, and at the center of this image is actually a black hole, but [17:28.040 --> 17:34.600] you can't see that because it's a black hole, and then on the background of this image there [17:34.600 --> 17:39.920] is an overlay of what's the optical domain, so not the radio domain from the Hubble Space [17:39.920 --> 17:45.440] Telescope, and this is what we have been able to capture with Lofar. [17:45.440 --> 17:49.800] So we're doing groundbreaking science, and we're going to do a lot more, we're in the [17:49.800 --> 17:54.840] middle of a big upgrade that's scheduled for the end of 2014, Vincisi is going to be replaced [17:54.840 --> 18:00.040] with Kavana, we're thinking about Allerta, but I've heard that Kavana has persisted [18:00.040 --> 18:05.880] alarms during, actually forced them today, so we might not need Allerta, and we're using [18:05.880 --> 18:07.440] Prometheus. [18:07.440 --> 18:11.440] We had this low band antennas, and the high band antennas, I briefly skimmed over that [18:11.440 --> 18:17.400] because the details, yeah, you have to cut some corners somewhere, but basically with [18:17.400 --> 18:21.960] Lofar 2 we'll be able to use both of them in a single observation. [18:21.960 --> 18:26.520] We'll also be able to use multiple beams, so we talked about the beam forming, currently [18:26.520 --> 18:32.120] Lofar is only able to have a single beam per observation, and we will also be able to point [18:32.120 --> 18:38.720] at different points in the sky, change that during an observation, and we call this mega [18:38.720 --> 18:45.800] mode, don't ask me why, yeah, we're completely refamping the skater system, we're now using [18:45.800 --> 18:51.600] Pytango, we have sufficiently upgraded hardware, Unibor 2s, we actually sell those to external [18:51.600 --> 18:57.200] institutes as well, so they're available, and we're drastically improving the timing [18:57.200 --> 19:02.000] distribution, so we're currently GPS based, everything is synchronized using GPS, all [19:02.000 --> 19:06.240] the stations across Europe, and we're going to use the white rabbit protocol that's made [19:06.240 --> 19:11.000] by CERN, that's based on a precision time protocol. [19:11.000 --> 19:15.780] Now very briefly this mega mode, what would this schematically look like, so this is basically [19:15.780 --> 19:21.200] what's running on cobalt or GPU cluster, and we do imaging and beam forming, and now we [19:21.200 --> 19:28.000] have one beam and several pointings, and they stay the same during observations, now we [19:28.000 --> 19:32.920] can have multiple beams, and we can repoint during the observation. [19:32.920 --> 19:36.600] That's going to create a lot of flexibility for the astronomers, and I'm going to be very [19:36.600 --> 19:40.560] excited with the science that is going to come from this. [19:40.560 --> 19:47.480] I want to leave you with some links, as mentioned our Astron repo, the Astron website, there's [19:47.480 --> 19:52.840] a very interesting page about 10 years of LOFAR, because we've actually existed, LOFAR [19:52.840 --> 19:57.400] has been in production since 2008, so that's been since quite some time, there's this very [19:57.400 --> 20:01.840] nice map on which you can actually see all the physical locations of all the stations, [20:01.840 --> 20:08.000] how many antennas are active or working or broken, so this is all open data, you can [20:08.000 --> 20:12.240] just look at this, and there's some history about all these presentations that I've done [20:12.240 --> 20:21.320] in the past, so any questions? [20:21.320 --> 20:47.640] Maybe first a short comment, the Chinese built a 500 meter dish, but what I really wanted [20:47.640 --> 20:52.800] to ask is whether you have collaboration with other astrophysical observations like square [20:52.800 --> 20:55.640] kilometer array or something like that? [20:55.640 --> 21:02.080] Well actually we collaborate on the square kilometer array, so there's definitely, can [21:02.080 --> 21:05.640] you repeat part of your question, because people were just leaving? [21:05.640 --> 21:10.360] Well, whether there is shared development in software and stuff? [21:10.360 --> 21:15.720] Yeah, yeah, for sure, for instance on CASACOR as I mentioned, but also WS Clean, we see [21:15.720 --> 21:24.240] contributions from external collaborators, and especially the jive, the part of jive [21:24.240 --> 21:32.200] that I showed at the very beginning, let me see, shouldn't be too long, so here I mentioned [21:32.200 --> 21:38.640] jive and EVN, this is a huge collaboration of parabolic dishes that are all connected [21:38.640 --> 21:49.400] and all the data is sent centrally to Dringelo at the headquarters of Astong, and that's [21:49.400 --> 21:57.000] actually where the EVN network processes all this data, but all these dishes that we use, [21:57.000 --> 22:00.600] those are not ours, right, those are from other parties. [22:00.600 --> 22:13.520] Someone's asking, since everything is processed digitally, can these telescopes focus on multiple [22:13.520 --> 22:17.960] targets at once by processing the data multiple times? [22:17.960 --> 22:22.960] That's an interesting question, and that depends, as I said you have the transient buffers which [22:22.960 --> 22:29.800] dump raw samples, but typically what we do is we already do beamforming on the station, [22:29.800 --> 22:33.040] and if you do the beamforming on the station, you're already looking at some point in the [22:33.040 --> 22:38.640] sky, you're only sending the result data from that beamforming to this cobalt cluster, [22:38.640 --> 22:43.480] you can't beamform again then, the data is lost, it's reductive in nature, but if you [22:43.480 --> 22:49.600] would send the raw sample data to cobalt, and it could somehow process all the data, [22:49.600 --> 22:56.240] which I don't think it has the bandwidth to do so, then you could, at a later point in [22:56.240 --> 23:02.000] time, point at any point in the sky again, which is, that's the job of the transient [23:02.000 --> 23:03.000] buffers. [23:03.000 --> 23:04.000] Thanks. [23:04.000 --> 23:17.800] Maybe I have a question here, would there be any ways or interests for amateur astronomers, [23:17.800 --> 23:27.240] or radio astronomers, to help or work with astronomy? [23:27.240 --> 23:31.920] Well there's definitely job positions on our page all the time, I think, I don't know [23:31.920 --> 23:36.480] if most are in the field of radio astronomy, but what typically happens, and I can briefly [23:36.480 --> 23:42.160] explain, is we have a system called Nordstar, in which astronomers submit their proposals [23:42.160 --> 23:47.040] and describe what they want to do with their instrument, and then we have a community that [23:47.040 --> 23:50.920] looks at that, and that actually accepts these proposals, and then they are scheduled. [23:50.920 --> 23:54.880] This is actually a very good question, because I completely skipped this in the talk, but [23:54.880 --> 23:59.840] I wanted to talk about this, and then things are scheduled in a system called TMS, and [23:59.840 --> 24:05.440] that basically looks at, okay, what part of these stations are required, and to do these [24:05.440 --> 24:10.680] observations and collect these data, then these data are collected and processed on cobalt [24:10.680 --> 24:14.800] and sap, and the data products are made available to these individual astronomers who've done [24:14.800 --> 24:20.040] that, and they get exclusive access for a period of time to do their research. [24:20.040 --> 24:21.040] Okay, thanks. [24:21.040 --> 24:31.280] I was more thinking about, just if someone is in Africa with a homemade dish, is there [24:31.280 --> 24:39.520] any way to capture something with an SDR, and add a little bit with data, or the scale [24:39.520 --> 24:42.360] of things is so different that... [24:42.360 --> 24:47.080] What's actually very important, or it's rather we need to model a lot of things and calibrate [24:47.080 --> 24:52.640] a lot of things, so that's why all the stations are roughly similar in shape, similar in shape, [24:52.640 --> 24:56.880] have similar hardware, so it would be definitely possible to buy your own station, build your [24:56.880 --> 25:02.720] own station, and have the same hardware, and then hook it up, that happens all the time. [25:02.720 --> 25:08.400] Different countries do that, buy stations, and then we add them, but having vastly different [25:08.400 --> 25:12.680] hardware and then hooking this up to the system would be very difficult, it's not designed [25:12.680 --> 25:13.680] to do that. [25:13.680 --> 25:21.720] Okay, so if you would make a very cheap station that could be built by amateur astronomers, [25:21.720 --> 25:27.960] you could deploy that everywhere in the world, and then make your public radio astronomy [25:27.960 --> 25:28.960] like that. [25:28.960 --> 25:39.560] Interesting, thanks.