Beyond FAIR - Talk for OOIFB Data Systems Committee

Transcript

1. What data infrastructure does open science need? B e y o n d FA I R Ryan Abernathey
 
 OOIFB Data Systems Committee Meeting, 2022
2. T h e O p e n S c i e n c e V i s i o n 2 https://earthdata.nasa.gov/esds/open-science for 👩🔬 in everyone_in_the_world: for 📄 in all_scientific_knowledge: 👩🔬.verify(📄) discovery = 👩🔬.extend(📄) This would transform the 🌎 by allowing all of humanity to participate in the scienti fi c process. What are the barriers to realizing this vision?
3. T h e O p e n S c i e n c e V i s i o n 3 https://earthdata.nasa.gov/esds/open-science for 👩🔬 in everyone_in_the_world: for 👨🔬 in everyone_else_in_the_world: 📄 = (👩🔬 + 👨🔬).collaborate() This would transform the 🌎 by allowing all of humanity to participate in the scienti fi c process. What are the barriers to realizing this vision?
4. ❤ 🥰 FA I R 🥰 ❤ 4 FAIR = Findable, Accessible, Interoperable, Reusable FAIR is great. Nobody disagrees with FAIR. But making data-intensive scienti fi c work fl ows FAIR is easier said than done. FAIR does not specify the protocols, technologies, or infrastructure. 
 
 FAIR is not a platform.
5. 5 Simulation Data-Intensive Science vs. ? https:// fi gshare.com/articles/ fi gure/Earth_Data_Cube/4822930/2 Equations Big Data Big Data 💡 Insights 
 💡 Understanding 
 💡 Predictions • known computational problem • optimized, scalable algorithm • standard architecture (HPC / MPI) • open-ended problem • exploratory analysis • “human in the loop” • ML model development & training • visualization needed • highly varied computational patterns / algorithms
6. • The word “platform” is terribly overloaded. • A platform is something you can build on—speci fi cally, new scienti fi c discoveries and new translational applications. Let’s call these projects. 📄 • For open science to take off at a global scale, everyone in the world needs access to the platform (like Facebook) • This is why we are excited about cloud, but cloud as-is (e.g. AWS) is not itself an open-science platform. • Does the open science platform need to be open? 🤔 C l a i m : O p e n S c i e n c e n e e d s a P l at f o r m 6 Infrastructure Platform Open- Science Project Platform Open- Science Project Open- Science Project Open- Science Project
7. O u t l i n e 7 10 mins The status quo of data-intensive scienti fi c infrastructure 10 mins Cloud computing and Pangeo 10 mins From Software to SaaS: Pangeo Forge and Earthmover 10 mins Where are things headed?
8. 8 Pa r t I : T h e S tat u s Q u o
9. D ata - I n t e n s i v e S c i e n c e I n f r a s t r u c t u r E : T h e S tat u s Q U O * 9 Personal Laptop Group Server Department Cluster Agency Supercomputer more storage, more CPU, more security, more constraints
10. S tat u s Q u o : W h at I n f r a s t r u c t u r e C a n W e R e ly o n ? 10 ✅ UNIX operating system 💾 ✅ Files / POSIX fi lesystems 🗂 ✅ Programming languages: C, FORTRAN, Python, R, Julia ✅ Terminal access ⚠ Batch queuing system ⏬ 
 HPC only ⚠ The internet 🌐 
 Not on HPC nodes! ⚠ Globus for fi le transfer 🔄 
 Not supported everywhere ❌ High level data services, APIs, etc. 
 Virtually unknown in my world
11. T h e “ D o w n l o a d ” M o d e l 11 a) file-based approach step 1 : dow nload step 2: clean / organize ` file file file b) database / api approach record file file file local disk query files step 3: analyze
12. 12 MB 😀 T h e “ D o w n l o a d ” M o d e l a) file-based approach step 1 : dow nload step 2: clean / organize ` file file file b) database / api approach record file file file local disk query files step 3: analyze
13. 13 GB 😐 T h e “ D o w n l o a d ” M o d e l a) file-based approach step 1 : dow nload step 2: clean / organize ` file file file b) database / api approach record file file file local disk query files step 3: analyze
14. 14 TB 😖 T h e “ D o w n l o a d ” M o d e l a) file-based approach step 1 : dow nload step 2: clean / organize ` file file file b) database / api approach record file file file local disk query files step 3: analyze
15. 15 PB 😱 T h e “ D o w n l o a d ” M o d e l a) file-based approach step 1 : dow nload step 2: clean / organize ` file file file b) database / api approach record file file file local disk query files step 3: analyze
16. P r i v i l e g e d I n s t i t u t i o n s c r e at e “ D ata F o r t r e s s e s * ” 16 Image credit: Moahim, CC BY-SA 4.0, via Wikimedia Commons *Coined by Chelle Gentemann
17. P r i v i l e g e d I n s t i t u t i o n s c r e at e “ D ata F o r t r e s s e s * ” 17 Image credit: Moahim, CC BY-SA 4.0, via Wikimedia Commons Data *Coined by Chelle Gentemann # step 1: open data open(“/some/random/files/on/my/cluster”) # step 2: do mind-blowing AI!
18. 🗂 Emphasis on fi les as a medium of data exchange creates lots of work for individual scientists (downloading, organizing, cleaning). Most fi le-based datasets are a mess— even simulation output. 😫 Yet the hard work of data wrangling is rarely collaborative. Outputs are not reusable and can’t be shared. 💰 Doing data-intensive science requires either expensive local infrastructure or access to a big agency supercomputer. This really limits participation. 🏰 Data intensive science is locked inside data fortresses. Limiting access to outsiders is a feature, not a bug. Restricts collaboration and reproducibility! ❄ Each fortress is a special snow fl ake. Code developed in one will not run inside another. p r o b l e m s w i t h t h e S tat u s Q u o 18
19. 19 Pa r t I I : C l o u d C o m p u t i n g a n d Pa n g e o
20. W h at a b o u t C l o u d ? 20 👩💻👨💻👩💻 Group A: 👩💻👨💻👩💻 Group B: 
 Research Education & Outreach Industry Partners *Coined by Fernando Perez Can we create a “data watering hole”* instead of a fortress?
21. O p t i o n A : V e r t i c a l ly i n t e g r at e d P l at f o r m 21 
 All the data 
 All the compute
22. O p t i o n B : I n t e r o p e r a b l e C l o u d - N at i v e D ata , S o f t w a r e , a n d S e r v i c e s 22 Data Provider’s Resources Data Consumer’s Resources Interactive Computing Community-Maintained ARCO Data Lake[s] Distributed Processing
23. Scientific users / use cases Open-source software libraries HPC and cloud infrastructure • Define science questions • Use software / infrastructure • Identify bugs / bottlenecks • Provide feedback to developers • Contribute widely the the open source scientific python ecosystem • Maintain / extend existing libraries, start new ones reluctantly • Solve integration challenges • Deploy interactive analysis environments • Curate analysis-ready datasets • Platform agnostic Agile development 👩💻 T h e Pa n g e o C o m m u n i t y P r o c e s s 23
24. A c c i d e n ta l P i v o t t o C l o u d 24
25. T h e Pa n g e o C l o u d - N aT i v e S ta c k 25 Cloud-optimized storage for multidimensional arrays. Flexible, general-purpose parallel computing framework. High-level API for analysis of multidimensional labelled arrays. Kubernetes Object Storage Rich interactive computing environment in the web browser. xgcm xr f xhistogram gcm-filters climpred Cloud Infrastructure Domain specific packages Etc.
26. A n a ly s i s - R e a d y, C l o u d O p t i m i z e d : 
 A R C O D ata 26 https://doi.org/10.1109/MCSE.2021.3059437 This also demonstrates the potential of the “hybrid cloud” model with OSN.
27. Pa n g e o h a s B r o a d A d o p t i o n 27
28. • Pangeo software can be deployed as a platform: 
 JupyterHub in the cloud with Xarray, Dask, etc., connected to ARCO data sources • But there are many distinct deployments of this platform - dozens of similar yet distinct JupyterHubs with different con fi gurations, environments, capabilities, etc 
 ➡ Sharing projects between these hubs is still very hard • Deploying hubs generally requires DevOps work (billed a developer time or contractor services). There is no “Pangeo as a Service” • Getting data into the cloud in ARCO format is hard and full of toil L i m i tat i o n s o f t h e Pa n g e o A p p r o a c h 28
29. • Non-agency scientists have many barriers to adopting cloud: 
 Overhead policies, purchasing challenges, lack of IT support, etc. • Cloud is too complicated! The services offered are not useful to scientists: 
 An extra layer of science-oriented services must be developed • Europe basically forbids scientists from using US-based cloud providers • Not much has changed for university scientists since 2017 C h a l l e n g e s w i t h C l o u d i n G e n e r a l 29
30. 30 Pa r t I I I : F r o m S o f t w a r e t o S a a S 
 
 Pa n g e o F o r g e & E a r t h m o v e r
31. T o o l s f o r C o l l a b o r at i o n 31 Some of the most impactful services used in open science…. These are all proprietary SaaS (Software as a Service) applications. 
 They may use open standards, but they are not open source. We (or our institutions) have no problem paying for them.
32. T h e M o d e r n D ata S ta c k 32 • In the past 5 years, a platform has emerged for enterprise data science called the Modern Data Stack • The MDS is centered around a “data lake” or “data warehouse” • Different platform elements are provided by different SaaS companies; integration through standards and APIs • No one in science uses any of this stuff https://continual.ai/post/the-modern-data-stack-ecosystem-fall-2021-edition
33. • Embrace commercial SaaS: a Modern Data Stack for Science • Cultivate community-operated SaaS: e.g. Wikipedia, Conda Forge, Binder, 2i2c Hubs, Pangeo Forge • We probably need a mix of both h o w c a n w e d e l i v e r a n o p e n s c i e n c e p l at f o r m i n a s c a l a b l e , s u s ta i n a b l e w ay ? 33 Community-operated SaaS for ETL (Extract / Transform / Load) of ARCO Data Our new startup. Building a commercial cloud data lake platform for scienti fi c data.
34. • Think in “datasets” not “data fi les” • No need for tedious homogenizing / cleaning steps • Curated and cataloged A R C O D ata 34 Analysis Ready, Cloud Optimzed ZRUN DERXWZHUHêVDWLVĆHGëRUEHWWHU  How a Data Scientist Spends Their Day +HUHèVZKHUHWKHSRSXODUYLHZRIGDWDVFLHQWLVWVGLYHUJHVSUHWW\VLJQLĆFDQWO\IURP ZHWKLQNRIGDWDVFLHQWLVWVEXLOGLQJDOJRULWKPVH[SORULQJGDWDDQGGRLQJSUHGLFWLY actually not what they spend most of their time doing, however. $V\RXFDQVHHIURPWKHFKDUWDERYHRXWRIHYHU\GDWDVFLHQWLVWVZHVXUYH\HG PRVWWLPHFOHDQLQJDQGRUJDQL]LQJGDWD<RXPD\KDYHKHDUGWKLVUHIHUUHGWRDVêG FRPSDUHGWRGLJLWDOMDQLWRUZRUN(YHU\WKLQJIURPOLVWYHULĆFDWLRQWRUHPRYLQJFRP Data scientist job satisfaction 60% 19% 9% 4% 5% 3% Building training sets: 3% Cleaning and organizing data: 60% Collecting data sets; 19% Mining data for patterns: 9% 5HĆQLQJDOJRULWKPV Other: 5% ,!;&!;!9$-'2ধ9;996'2&;,'1 How do data scientists spend their time? Crowd fl ower Data Science Report (2016) What is “Analysis Ready”?
35. • Compatible with object storage 
 (access via HTTP) • Supports lazy access and intelligent subsetting • Integrates with high-level analysis libraries and distributed frameworks A R C O D ata 35 Analysis Ready, Cloud Optimzed What is “Cloud Optimized”?
36. A R C o D ata i s Fa s t ! 36 https://doi.org/10.1109/MCSE.2021.3059437 This also demonstrates the potential of the “hybrid cloud” model with OSN.
37. P r o b l e m : 37 Making ARCO Data is Hard! Domain Expertise: 
 How to fi nd, clean, and homogenize data Tech Knowledge: 
 How to ef fi ciently produce cloud-optimized formats Compute Resources: 
 A place where to stage and upload the ARCO data Analysis Skills: 
 To validate and make use of the ARCO data. To produce useful ARCO data, you must have: Data Scientist 😩
38. W h o s e J o b i s i t t o M a k e A R C O D ata? 38 Data providers are concerned with preservation and archival quality. Scientists users know what they need to make the data analysis-ready.
39. Pa n g e o F o r g e 39 Let’s democratize the production of ARCO data! Domain Expertise: 
 How to fi nd, clean, and homogenize data 🤓 Data Scientist
40. I n s p i r at i o n : C o n d a F o r g e 40
41. 41 Pangeo Forge Recipes Pangeo Forge Cloud Open source python package for describing and running data pipelines (“recipes”) Cloud platform for automatically executing recipes stored in GitHub repos. https://github.com/pangeo-forge/pangeo-forge-recipes https://pangeo-forge.org/
42. Pa n g e o F o r g e R e c i p e s 42 FilePattern StorageConfig Recipe Executor Describes where to fi nd the source fi les which are the inputs to the recipe Describes where to store the outputs of our recipe A complete, self- contained representation of the pipeline Knows how to run the recipe. https://pangeo-forge.readthedocs.io/
43. Pa n g e o F o r g e C l o u d 43 Feedstock Contains the code and metadata for one or more Recipes Bakery https://pangeo-forge.org/ Storage Runs the recipes in the cloud using elastic scaling clusters GCS
44. V i s i o n : C o l l a b o r at i v e D ata C u r at i o n 44 Feedstock 🤓 Data User 🤓 Data Producer 🤓 Data Manager These data look weird… …Oh the metadata need an update. Ok I’ll make a PR to the recipe.
45. 45 Oceanographers building a full-stack cloud SaaS automation platform. https://twitter.com/Colinoscopy/status/1255890780641689601 Charles Stern
46. 🙌 Pangeo Forge Cloud is live and open for business! 
 pangeo-forge.org 💣 Our recipes often contain 10000+ tasks. We are hitting the limits on Prefect as a work fl ow engine. Currently refactoring to move to Apache Beam. 😫 Data has lots of edge cases! This is really hard. 🌎 But we remain very excited about the potential of Pangeo Forge to transform how scientists interact with data. C u r r e n t s tAT U S 46
47. Earthmover Founders: Ryan Abernathey Joe Hamman Product: ArrayLake • High performance for analytics (based on Zarr data model) • Ingest and index data from archival formats (NetCDF, HDF, GRIB, etc.) • Automatic background optimizations • Versioning / snapshots / time travel • Data Governance • Compare to Databricks, Snow fl ake, Dremio Mission: To empower people to use scientific data to solve humanity’s greatest challenges A Public Benefit Corporation
48. 48 Pa r t I V : W h e r e a r e w e H e a d i n g ?
49. compute node P i l l a r s o f C l o u d N at i v e S c i e n t i f i c D ata A n a ly t i c s 49 1. Analysis-Ready, Cloud-Optimized Data 2. Data-Proximate Computing 3. Elastic Distributed Processing compute node compute node compute node compute node compute node compute node compute node compute node compute node Compute Environment
50. S e pa r at i o n o f S t o r a g e a n d C o m p u t e 50 Storage costs are steady. Data provider pays for storage costs. May be subsidized by cloud provider.
 (Thanks AWS, GCP, Azure! Or can live outside the cloud (e.g. Wasabi, OSN) Compute costs for interactive data analysis are bursty. Can take advantage of spot pricing Multi-tenancy:
 We can all use the same stack, but each institution pays for its own users. This is completely di ff erent from the status quo infrastructure!
51. O p e n S c i e n c e P l at f o r m 51 “Analysis Ready, Cloud Optimized Data” 
 Cleaned, curated open-access datasets available via high-performance globally available strorage system “Elastic Scaling” 
 Automatically provision many computers on demand to accelerate big data processing. “Data Proximate Computing” 
 Bring analysis to the data using any open-source data science language. Generic cloud object storage Generic cloud computing Data Library Compute Environment Expert Analyst Direct Access via Jupyter Non-Technical User Access via apps / dashboards / etc. Runs on any modern cloud-like platform or on premises data center Web front ends Downstream third-party services / applications
52. F e d e r at e d , E x t e n s i b l e M o d e l 52 Compute Environment Data Library Data Library Data Library Data Library Data Library Compute Environment Compute Environment Front-end Services
53. D ata G r av i t y 53 “Data gravity is the ability of a body of data to attract applications, services and other data." - Dave McCrory 
 
 
 
 
 NASA (200 PB) NOAA BDP ASDI (incl. CMIP6) 
 NCAR Datasets 
 etc… 
 
 Planetary Computer 
 NOAA BDP 
 
 
 
 Earth Engine 
 NOAA BDP Descartes 
 Pangeo 
 
 
 SentinelHub Climate Change Atmosphere Marine ECMWF DOE XSEDE HECC NCAR
54. D ata G r av i t y 54 What is the stable steady-state solution? DOE XSEDE HECC NCAR ?
55. W e n e e d a 
 g l o b a l S c i e n t i f i c D ata C o m m o n s 55 Need to be exploring: edge storage, decentralized web, web3 DOE XSEDE HECC NCAR ?
56. S h o u t O u t s 56
57. • What’s the right model to deliver data and computing services to the research community? Commercial vendors? Co-ops? • How can we avoid recreating existing silos in the cloud? • Who should we pay for cloud infrastructure for the science community? University? Agency? PI? • How can we make cloud interoperate more with HPC and on-premises computing resources? D i s c u s s i o n Q u e s t i o n s 57
58. L e a r n M o r e 58 http://pangeo.io https://discourse.pangeo.io/ https://github.com/pangeo-data/ https://medium.com/pangeo @pangeo_data