4/16/2018. Building the HASSP Or: Crazy Ideas Are Okay

Size: px

Start display at page:

Download "4/16/2018. Building the HASSP Or: Crazy Ideas Are Okay"

Theodora Dorsey
5 years ago
Views:

1 4/16/2018 Building the HASSP Or: Crazy Ideas Are Okay

2 About me I m a Senior SQL Server DBA at IGS Energy in Dublin, Ohio. I ve been using SQL Server since SQL Server 2000 and I love it. I also enjoy DevOps, release management, and PowerShell. Contact Info 4/16/2018 2

3 I want to tell you a story This isn t going to be like most sessions you might attend today But I promise it will be interesting And it does involve SQL Server but is that what it s truly all about? 4/16/2018 3

4 It s 2016, and I m bored Sure, I m working full time I love what I do (especially working with the Data Platform) I love where I work I love who I work with I m also out in the community Speaking Teaching Blogging But something s missing 4/16/2018 4

5 Then something amazing happened Microsoft Announces SQL Server for Linux! Twitter loses it mind Everyone and their brother installs and starts trying it out Suddenly everyone wants to know more about Linux 4/16/2018 5

6 And that got me thinking People were coming up for all kinds of use cases for SQL Server on Linux But it was still in public preview Not like people were going to jump to production right away Wouldn t it be cool to use this a showcase for the Data Platform? 4/16/2018 6

7 Linux is pretty rad as an OS It s free It s got a small footprint It uses less resources than a Windows-based machine It can run on a toaster or refrigerator (no, really) 4/16/2018 7

8 So where do we want to put a SQL Server? Part publicity stunt, part actual practical application At the end of the day, I don t want just a lame gimmick I wanted a project that could collect and store data for either retrieval later or able to be queried remotely 4/16/2018 8

9 Good artists copy I had been following a few science blogs and I was reading about how amateurs were launching high altitude balloons. Everyone from boy scouts, 7 th grade science students, and way more serious people were doing it. It looked fun and adventurous! So 4/16/2018 9

10 HASSP-1 was born I figured these balloons carried other payloads, why not a SQL Server? So I stood up in the middle of our open office, and asked Who wants to put a SQL Server in space? A lot of people laughed But not everyone 4/16/

11 So I formed a team We sat down and talked through the challenges of doing this project. Here were the major areas of concern: What kind of balloon flies that high? Do we need permission (is anyone liable to go to jail)? What kind of data would we want to collect while it flies? How do we get that data? How we find the balloon when it lands? How we ensure it survives the flight? And oh yeah, and how do we actually fit (and power) a SQL Server on a balloon payload? How much is this going to cost? 4/16/

12 We solved some problems We re not going to go to jail if we follow rules We learned about how HABs work We chose some data points which enabled us to decide what kind of sensors to buy which led to me burning myself with a soldering iron a lot We ran some simulations We decided to use APRS to track the HAB 4/16/

13 Problem 1: What goes up High altitude ballooning is nothing new, and, really, not that hard if you think about it You fill a balloon up with helium, and you let it go Except not all balloons, or helium, are equal You need a latex balloon designed for atmospheric flight You also need high-grade helium, and you might need a lot of it Balloon size helps determine your max ceiling, your flight time, and total lift capacity Bigger balloons = more helium Fun fact! You can t return the helium you don t use for a refund Wound up with a 1200g balloon from High Altitude Science And other accessories, like a lift scale and inflator kit 4/16/

14 Problem 2: Are we going to go to jail? Launching high altitude balloons is totally legal, and depending on your total payload weight, you don t even need permission! However, there a few things to know about to make your life (and the lives of those around you) easier Launch sites Radar reflectors Payload weights NOTAMs ATC communication 4/16/

15 Problem 3: What do we want to collect? Since this thing is going up pretty high, we figured we d want to collect some atmospheric data I looked for some cheap consumer sensors for: Temperature Pressure Humidity After looking at different models, we settled on three sensors: Sensor 1 Sensor 2 Sensor 3 4/16/

16 Turns out that soldering irons are hot This was the first big hurdle for the project: how can we create a circuit capable of detecting this data? Creating simple circuits and testing them with Arduino is pretty simple, and fun! I developed a proof-of-concept circuit using Arduino to show the sensors worked by writing data to consoles So when the time came, we could (should?) be able to connect these to a similar system and have the data written to SQL Server (right?) 4/16/

17 Problem 4: How do we find this when (if) it lands? One of the requirements of sending balloons up is that you need a way to track and report on position data One really good way to do this is with amateur radio (Ham Radio) You can use the APRS network to transmit data packets One big thing about this, though: you need a license (and I didn t have one) 4/16/

18 Just call me KE8SQL Yes, I actually studied for and obtained my Technician Class license With that done, it was time to learn about implementing APRS APRS = Automatic Packet Reporting System VHF on Data packets are transmitted from an object Packets usually contain position and telemetry info (and other pieces of info, depending on the transmitter capabilities) These packets are detected and re-transmitted by repeater stations These stations also forward these packets to internet gateways to users to view them on maps Like 4/16/

19 I need a volunteer! Let s see APRS in action! 4/16/

20 Our first real equipment purchase! I picked up an APRS on-board transmitter with 2M dipole antenna for the project Created by a company called Byonics Designed specifically for HAB projects Has on-board GPS receiver and sends packets containing location and vector information Battery powered 1W output 4/16/

21 Predicting the unpredictable? The tracker is cool for knowing where it is How do you figure out where your balloon is going to go? And what about when the balloon bursts? There are predictors out there to help: Predictor 1 Predictor 2 Predictor 3 4/16/

22 How we attach a SQL Server to a balloon? We ve got a bunch of sensors that can collect data We ve got a way to track the balloon in flight We ve got a rough idea of the rules and regulations around launching the balloon So, uh, how do we get it to run SQL Server? 4/16/

23 Step 1: Finding a server Since space and weight were at a premium, we needed something to could run SQL Server and still weigh next to nothing A lot of people want to yell Raspberry Pi! Nope; ARM!= x64 SOC Solutions We first looked at a Latte Panda: A 64-bit platform with built-in Arduino support But we settled on an Intel Joule 4/16/

24 System Requirements SQL Server has system requirements, so that seems like a pretty good place to start: 64-bit processor, at least 1.4GHz 4GB of RAM 6GB of storage for binaries and system components 4/16/

25 Meet the Joule! 1.7 GHz Quad-Core 4GB RAM 16GB On-board storage (expandable) 24x48mm* Bluetooth and WiFi USB3 4K Display Capable GPIO Headers! Great news! It s discontinued 4/16/

26 Step 2: Building a home for all the components We bought a bunch of ice cream from Graeter s It comes in a super-duper efficient Styrofoam container Some glue, balsa wood, nails, and zip ties later How do we make it survivable inside the unit? 4/16/

27 Step 3: Oh yeah, software SQL Server runs on the hardware We know the sensors are wired up correctly How we do make them talk to each other? Initially, wanted to use GPIO on the Joule TERRIBLE documentation Ended up using Johnny-Five, a robotics & IoT platform Built on JavaScript Shell scripts auto-start when the power is engaged Polls sensors and writes their data SQL Server 4/16/

28 Step 4: Rounding out the build Parachute (kind of important) HVAC Cameras Lines and knots 4/16/

29 Hi, I m the HASSP module. Nice to meet you! Here s the module! 4/16/

30 Launch planning Where? When? How? 4/16/

31 Where should we launch? Obviously wanted some place out of the way Where we launched had a big factor on where the payload would travel, both on ascent and descent After running a bunch of simulations, we decided on an airfield near New Knoxville, OH Neil Armstrong Airport (AXV) Launching at an airport had other advantages, too: Easier to work with FAA for paperwork 4/16/

32 When should we launch? Earlier, and not because we didn t want to completely lose a day We initially decided on a Tuesday morning Less traffic on the roads to track and follow the balloon More eyeballs for viewing Weather played an important factor, too 4/16/

33 How should we launch? Sounds easy, right? Fill up a balloon, tie the payload, and whoosh, let it go? Slightly more complicated: Turn on and test sensor suite and server Attach and start cameras Put in heating pads Turn on the tracker and validate the squawk Payload assembly Train assembly All outdoors (which came back to haunt us) with a running clock 4/16/

34 Launch Day! HASSP 1 was scheduled to launch on Tuesday, July 11 th at 8:30 AM We thought we were ready NOTAMs filed Gear all checked out Team was HYPED! Ground and temperature testing had gone well Our live stream system was set up and ready to go 4/16/

35 Launch Day? There had been severe storms the day before our scheduled launch, and storms predicted later in the day as well We had a decision to make and we pushed ahead We arrived at the airport to high winds and things went downhill from there 4/16/

36 So what do you think happened 4/16/

37 Well, here s what happened 4/16/

38 This guy? Not really happy 4/16/

39 What data did we get? 4/16/

40 What just happened? We thought we were ready we weren t ready A lot of IT parallels here! We didn t do enough testing We made bad assumptions We should have slowed down instead of speeding up We didn t have the right tools 4/16/

41 What we learned Setup time could be better (a lot of wasted time) Pre-staging some components beforehand? Our cameras weren t that great We need a better way to measure helium input and lift We need a better place to work Rural Wi-Fi isn t quite the best Our live streaming system wasn t ready for prime-time Budget? What budget? 4/16/

42 But there were some positives! SQL Server DID fly that day! And it recorded data! We proved this thing might actually work 4/16/

43 Back to the drawing board We had a team retro We talked about what worked, and what didn t We came up with deliverables to make things better Improved internal construction of the payload (better heat management) The search for better cameras How can we accurately measure lift? How can we be more efficient on the ground? 4/16/

44 HASSP-2: The SeQueL We made some significant improvements Better live-streaming set up Cellular WiFi Improved microphone Better streaming setup back home We worked with the airport authority to do our work in a hanger, out of the elements We pre-tied a bunch of knots and lines prior to heading to the launch site We used a fish scale to help get better accurate readings on lift We got some new cameras (thanks Brent Ozar!) 4/16/

45 So how d we do? Armed with experience and a better plan, we set out again for the airport 4/16/

flight estimations but how close would they stick to them?

46 And then, we waited Our estimated flight time was ~2.5 hours 2 hours to reach maximum altitude 30 minutes to fall back to earth We had flight estimations but how close would they stick to them? Actually, pretty darn close! The predictors had various landing predictions in and around Urbana, Ohio (about 80 miles away) Turns out 4/16/

47 Recovery Everything was going great: we were tracking the progress just fine However, once the transmitter loses line of sight to repeaters, we lose reporting from APRS listeners If we can get close enough, my HT could do the trick 4/16/

48 And we got it all on tape! We had two cameras One facing straight out Once facing down Both published to YouTube Let s see some highlights! 4/16/

49 Mission accomplished? We successfully launched a balloon to > 100,000! We tracked it! We recovered it! Wait, wasn t there a SQL Server somewhere in this project? SQL Server did run for part of the flight but the server shut down and we only got about 30 minutes of data 4/16/

50 HASSP-3? So, are we going to go again? You bet! New SOC coming soon! Better system and power protection? Real-time querying? Follow along on twitter or my website for announcements About that budget 4/16/

51 The REAL lesson of HASSP Yeah, this used SQL Server Yeah, it required some unique problem solving Yeah, it required learning a bunch of new skills The real breakthrough though, was getting to know and work with people I work with every day to do something extraordinary Don t be afraid to something crazy It could be a side project Or maybe a different way of tackling a project 4/16/

52 I couldn t (and wouldn t have wanted) to do this alone Special thanks to the team! Gene Pavell and Matt Keller Capsule Construction Paul Bahler Software Engineering Robert Schmidt and Jacob Sturm Live Streaming Sarah Kenny Marketing and Production Julie Dorazio Data Modeling and Database Design Rick Kierner Launch Site Planning and Coordination Ryan Matthews Flight Planning and Regulatory Affairs Chris Klein Logistics and Hardware Support, Photography 4/16/

53 Special thanks to Brent Ozar Camera Donations IGS Support and Encouragement The SQL Community for following along! 4/16/

54 You can do this, too! HAB References FAA Part 101: FCC : 4/16/

55 About me I m a Senior SQL Server DBA at IGS Energy in Dublin, Ohio. I ve been using SQL Server since SQL Server 2000 and I love it. I also enjoy DevOps, release management, and PowerShell. Contact Info 4/16/

AC : REAL-TIME VIDEO TRANSMISSION FROM HIGH ALTITUDE BALLOON: AN INTERDISCIPLINARY SENIOR DESIGN PROJECT

AC : REAL-TIME VIDEO TRANSMISSION FROM HIGH ALTITUDE BALLOON: AN INTERDISCIPLINARY SENIOR DESIGN PROJECT AC 2010-1120: REAL-TIME VIDEO TRANSMISSION FROM HIGH ALTITUDE BALLOON: AN INTERDISCIPLINARY SENIOR DESIGN PROJECT Brent Guenther, Wright State University Bruce Rahn, Wright State University Mark Falknor,