SlideShare a Scribd company logo
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
Haden Land
haden.land@lmco.com
Jason Loveland
jason.c.loveland@lmco.com
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Orion Program Overview
 High Data Rate Ingest Use Case
 MACH-5 INSIGHT℠ Platform
 SFDU and Protobuf
 HBase bulk loading using Spark Streaming
 TRACE Visualization
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
Orion Multi-Purpose Crew Vehicle, NASA’s first spacecraft designed for
long-duration, human-rated deep space exploration. Orion will transport
humans to interplanetary destinations beyond low Earth orbit, such as
asteroids, the moon and eventually Mars, and return them safely back to
Earth.
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Carry out a robust human and robotic
exploration program that is both sustainable
and affordable
 Extend human presence across the solar
system to our planets, asteroids and other
destinations
 Develop the innovative technology,
knowledge, and infrastructure needed to
support more challenging human space
exploration missions
 Promote international and commercial
participation in exploration to further advance
U.S. scientific, security and economic
interests
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Uncrewed Orion launched from Cape Canaveral Air Force Base on a Delta IV Heavy
rocket
 Orbited the Earth twice, traveling approximately 3,600 miles above the Earth’s
surface
• 15 times farther than the International Space Station.
• Generated more than 80% of the return velocity experienced during a reentry from the moon, which allows engineers
to model expected reentries from future missions in deep space.
 Orion travelled twice through the Van Allen Radiation Belt, a dense radiation field of
highly energetic charged particles surrounding Earth
• This exposure will help the team understand and mitigate radiation exposure, and to develop protective solutions
before the first crewed mission.
 At the conclusion of EFT-1, Orion reentered Earth’s atmosphere at 20,000 mph and
enduring temperatures of up to 4,000° Fahrenheit
 After the parachutes deployed, the spacecraft slowed to 20 mph splashing down in
the Pacific Ocean, near the Baja Peninsula.
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
• How much telemetry is there, anyway? Is it “big
data”?
• Orion is transitioning to Big Data
• LOTS! PB+. On board, three 1 Gb networks are kept full
• Volume, velocity, variety
• What is Spacecraft Telemetry?
• Sensor measurements to monitor everything
• Microsecond Timestamps
• 12 discrete time sources
• More complex time series
• Human Space Flight requires fault tolerance
• What produces telemetry?
• For Orion, the telemetry is produced in a variety of
simulation and test environments which includes at least
7 differ labs across the US.
• How many items are there to measure in Orion?
• The XTCE representing the flight software telemetry has
over 3,000,000 elements, of which over 300,000 are
related to currently taken measurements. There are
many more when considering the ground system and
other subsystems.
• How is telemetry analyzed today?
• Traditionally, the responsible engineer
watches a handful of the telemetry
indicators on the real time monitors as a
test progresses, or in review. Certain
behavior and reporting will result in
further analysis on a few other
measurands.
• Most values of most telemetry
measurements are ignored if they are
not out of limits
• Specific studies are done on some
measurements in historical context,
usually after detection of anomalous
behavior, to determine if the behavior
has been observed in the past.
• Automation has been applied to the
problems of automatically collecting
data, automatically running scripted
tests, and automatically detecting out-of-
limit values
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
Orion’s next mission (EM-1) in 2018
 2 weeks instead of 4 hours
 4 times as many computers
 Twice as many instruments
 Subsystems that support Human Flight!
EM-1 will be orders of magnitude bigger and more
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Technologists/Platform Developers
• Networking
• Disk storage strategies
• Modern languages and techniques: Scala, R
• Big Data tools: Hadoop ecosystem, NoSQL databases, Stream and Batch Processing
 Analytics /Data Scientist(Ponderers)
• Tools of the trade
• Visualization & Communication
• Data Science
 Statistics
 Design of experiments
 Statistical tools: SAS, SPSS, R, MATLAB
 Modeling
 Data Mining
 Linear algebra
 Data cleansing, sampling, warehousing
 Subject Matter Experts
• System and Subsystem Experts
• Program Tools Experts
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
Horizontally scalable, fault tolerant prototype for processing High Data
Rate telemetry from Orion Test Rigs including ingest, processing,
transform, storage, advanced analytics
 Satisfy Data Rate: ~300MB/s (97MB/s X 3 planes)
 Provide common data wire protocol that is:
 Flexible, expressive and compact
 Programming language agnostic
 Provide system to scale telemetry ingest from Test Rigs to Mach-5
within technical performance metrics
 Provide solution that is continuously measurable and monitorable
 Provide Advanced Analytics to identify anomalies on ingest
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Enables Orion to teach us lessons about it’s own experience leveraging data
to tell the story
 Reduces time to develop and test Orion by identifying failures early!
 Uses data points from one mission to help predict if mission critical
components will succeed or fail in specific future mission circumstances
 Permits queries to be retrieved across historical batch and live streaming
views within seconds
 Streams historical data back as if it were live
 Overlays data across tests and missions
 Stores and processes much larger data sets than any other space craft in
history
Scales to meet the needs of future missions to
keep the conversation all the way to Mars
Data Reader/Simulator Ingest
Packet
Measurands
(GPBs)
Kafka
Message
Bus
Packet
Measurands
(GPBs)
Deduplicat
ion (Spark)
HBase
Writer
(Spark)
mach5-
sample ObjSplitter +
Decom
(GDS)
C++ client Reads
Packets and
Decommutates
Tlm
Data
Packet Measurands
GPB File
(represents a Packet(s)
and contains
decommutated
measurands)
Header Metadata
apid:seqctr:time: value1
…
..
apid:seqctr:time: valueN
mach5-
sample
(Spark)
Packet
Measurands
(GPBs)
Lockheed Martin Proprietary Information
Storage
Analytics
HDFS
HFiles
(HBase-RDD)
Mach-5 Data Ingest for Orion
HBase
Web/UITomcat
Glassfish
Etc.
Trace
FOSS
widgets
Aggregatio
n
(Spark)
Alerting
(Spark)
Limit
Checking
(Spark)
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
• Leverage space standards and wrapping
modern data serialization technologies
• SFDU (Standard Formatted Data Unit)
• CCSDS Standard
• Consultative Committee for Space
Data Systems – Founded 1982
• Delimiting data and providing an
internal index
• Transport and archival format
• Provides ease of integration with space
systems
• Should be usable 25+ years from now
• SFDU wraps serialized data, in our case
Protobuf
CCSD Z 001L
LMCO U 009L
V
V
LMCO S 010
LMCO I 010
L
L
V
V
Header
Body
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Protocol Buffers
• Compact, compressible
• Many language bindings: Java, C++, Python, etc…
• Self describing schema, auto-generates code
• Compared to other Serialization libraries (i.e. Avro), allows for
greater data type expression.
 Ex: Can natively support unsigned Long and Unsigned Integer
on C++ side
 Allows for cleaner generated code bindings
• Widely used within the Hadoop ecosystem
• Widely used within Lockheed Martin telemetry processing
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Enables Orion to scale data architecture linearly on
FOSS
 Allows for multiple test rigs to simultaneously write
data to be ingested at high data rates
 Provides ability for analytics processing to
independently pull and push data
 Provides replay capabilities for guaranteed, exactly
once processing semantics per partition
 Provides fault tolerance and replication
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Enables Orion to scale data architecture linearly on FOSS
 Real-time Analytics reduce time to develop and test the Orion by identifying
anomalies early!
 Allows for analytics to be run in batch and streaming mode
 Analytics libraries that scale
 Integration with Kafka to independently pull and push data
 Runs on YARN to reduce complexity of managing resources
 Provides guaranteed, exactly once processing semantics per partition using
Kafka Direct Connect
 Provides fault tolerance and check pointing
 Integration with HBase (SparkOnHBase)
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
Pull Data off of Kafka and Transform from
SFDU to Protobuf HBase Column Format
•Data bucketed to be aligned to appropriate Region. (HBase
table is pre-split). Handled by Spark on Yarn
Sort Rows and Partitioning by HBase
Regions
•Sort happens locally
•HFiles are guaranteed to contain all rows for a specific region
•This avoids CPU intensive work by HBase
Construct HFiles Write to HDFS
•HFiles are written to HDFS
•Writing HFiles is much faster than HBase puts
Bulk Load into HBase
•Bulk Load is called on each HFile
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Data arrives in random order and HBase
automatically sorts data by time
 Enables Orion to scale data architecture linearly on
FOSS
 Provides random access to all telemetry with low
latency access
 Provides data store to perform batch processing
using Mapreduce and Spark
 Custom filters
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Single-Column Schema
• Single Colum is packed with every piece of metadata for a row. Numerous serialization formats can be
used: json, protobuf, avro, etc…
 Benefits
• Compact, greatly decreases row footprint
• Higher write throughput. Under the hood HBase performs a Put for each column of a row. We go from ‘N’
puts in a multi column schema to 1
• Caching for queries
 Detractors
• Increased overhead to update the column
• If your column schema has the potential to change, could end up with multiple versions of the serialized
column value
• Slower reads because requires that all of the data be returned and deserialized
• Requires a custom HBase filter to filter on individual members of the column value
• Integration with 3rd Party BI tools and the rest of the Hadoop ecosystem
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
Detect abnormal activity during a test and notify
operators early as possible. (Tests can run for
weeks!)
 Provide limit checking on individual
measurements
 Combine measurements to compute higher
level understanding (i.e. momentum
calculation)
 Aggregation and statistical summaries
 Requirements validation
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
 Analytics!
 We need the help of the
community to develop effective
data platforms and analytics that
can scale with Orion missions.
© 2015 Lockheed Martin Corporation
LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in
the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation

More Related Content

Viewers also liked

Solution Workshop – I Implemented Contracts – Now What?
Solution Workshop – I Implemented Contracts – Now What?Solution Workshop – I Implemented Contracts – Now What?
Solution Workshop – I Implemented Contracts – Now What?
SAP Ariba
 
Praktikum ii mandel 1 ( ma)
Praktikum ii mandel 1 ( ma)Praktikum ii mandel 1 ( ma)
Praktikum ii mandel 1 ( ma)
aris trea
 
latest catalogue
latest cataloguelatest catalogue
latest catalogue志珍 范
 
Network Growth Strategies: Getting More Out of Your Ariba Investment
Network Growth Strategies: Getting More Out of Your Ariba InvestmentNetwork Growth Strategies: Getting More Out of Your Ariba Investment
Network Growth Strategies: Getting More Out of Your Ariba Investment
SAP Ariba
 
Introducing Oracle Advanced Financial Controls Cloud Service
Introducing Oracle Advanced Financial Controls Cloud ServiceIntroducing Oracle Advanced Financial Controls Cloud Service
Introducing Oracle Advanced Financial Controls Cloud Service
Dane Roberts
 
Finit solutions - The Benefits of a Rolling Forecast
Finit solutions - The Benefits of a Rolling ForecastFinit solutions - The Benefits of a Rolling Forecast
Finit solutions - The Benefits of a Rolling Forecast
finitsolutions
 
Service Entry Sheets for Major and Complex Capital Projects
Service Entry Sheets for Major and Complex Capital ProjectsService Entry Sheets for Major and Complex Capital Projects
Service Entry Sheets for Major and Complex Capital Projects
SAP Ariba
 
Managing Complex Services in SAP and SAP Ariba from a Client Perspective
Managing Complex Services in SAP and SAP Ariba from a Client PerspectiveManaging Complex Services in SAP and SAP Ariba from a Client Perspective
Managing Complex Services in SAP and SAP Ariba from a Client Perspective
SAP Ariba
 
Become Jythonic in FDMEE (KSCOPE15)
Become Jythonic in FDMEE (KSCOPE15)Become Jythonic in FDMEE (KSCOPE15)
Become Jythonic in FDMEE (KSCOPE15)
Francisco Amores
 
Planning Process Considerations
Planning Process ConsiderationsPlanning Process Considerations
Planning Process Considerations
finitsolutions
 
How to play Golf on the first day of Go-Live for a JD Edwards EnterpriseOne ...
 How to play Golf on the first day of Go-Live for a JD Edwards EnterpriseOne ... How to play Golf on the first day of Go-Live for a JD Edwards EnterpriseOne ...
How to play Golf on the first day of Go-Live for a JD Edwards EnterpriseOne ...
KPIT
 
Getting the Most Out of EPM: A deep dive into Account Reconciliation Manager
Getting the Most Out of EPM: A deep dive into Account Reconciliation ManagerGetting the Most Out of EPM: A deep dive into Account Reconciliation Manager
Getting the Most Out of EPM: A deep dive into Account Reconciliation Manager
finitsolutions
 
Introduction to Oracle Hyperion Planning - New Features in 11.1.2.4
Introduction to Oracle Hyperion Planning - New Features in 11.1.2.4Introduction to Oracle Hyperion Planning - New Features in 11.1.2.4
Introduction to Oracle Hyperion Planning - New Features in 11.1.2.4
finitsolutions
 
UGC NET Classroom coaching photographs
UGC NET Classroom coaching photographsUGC NET Classroom coaching photographs
UGC NET Classroom coaching photographs
UGC NET Astral Education
 
Oracle Data Integration - Overview
Oracle Data Integration - OverviewOracle Data Integration - Overview
Oracle Data Integration - Overview
Jeffrey T. Pollock
 
Unit 1 Overview of International Business
Unit 1 Overview of International BusinessUnit 1 Overview of International Business
Unit 1 Overview of International Business
Charu Rastogi
 
Oracle Architecture
Oracle ArchitectureOracle Architecture
Oracle Architecture
Neeraj Singh
 
Literal e texto-integro_de_contratos_colectivos_vigentes_oct_2015
Literal e texto-integro_de_contratos_colectivos_vigentes_oct_2015Literal e texto-integro_de_contratos_colectivos_vigentes_oct_2015
Literal e texto-integro_de_contratos_colectivos_vigentes_oct_2015
Gobierno Autónomo Descentralizado de la Parroquia Rural de Puembo
 

Viewers also liked (20)

PowerPoint Presentation
PowerPoint PresentationPowerPoint Presentation
PowerPoint Presentation
 
Solution Workshop – I Implemented Contracts – Now What?
Solution Workshop – I Implemented Contracts – Now What?Solution Workshop – I Implemented Contracts – Now What?
Solution Workshop – I Implemented Contracts – Now What?
 
Executive Severance Chapter 1
Executive Severance Chapter 1Executive Severance Chapter 1
Executive Severance Chapter 1
 
Praktikum ii mandel 1 ( ma)
Praktikum ii mandel 1 ( ma)Praktikum ii mandel 1 ( ma)
Praktikum ii mandel 1 ( ma)
 
latest catalogue
latest cataloguelatest catalogue
latest catalogue
 
Network Growth Strategies: Getting More Out of Your Ariba Investment
Network Growth Strategies: Getting More Out of Your Ariba InvestmentNetwork Growth Strategies: Getting More Out of Your Ariba Investment
Network Growth Strategies: Getting More Out of Your Ariba Investment
 
Introducing Oracle Advanced Financial Controls Cloud Service
Introducing Oracle Advanced Financial Controls Cloud ServiceIntroducing Oracle Advanced Financial Controls Cloud Service
Introducing Oracle Advanced Financial Controls Cloud Service
 
Finit solutions - The Benefits of a Rolling Forecast
Finit solutions - The Benefits of a Rolling ForecastFinit solutions - The Benefits of a Rolling Forecast
Finit solutions - The Benefits of a Rolling Forecast
 
Service Entry Sheets for Major and Complex Capital Projects
Service Entry Sheets for Major and Complex Capital ProjectsService Entry Sheets for Major and Complex Capital Projects
Service Entry Sheets for Major and Complex Capital Projects
 
Managing Complex Services in SAP and SAP Ariba from a Client Perspective
Managing Complex Services in SAP and SAP Ariba from a Client PerspectiveManaging Complex Services in SAP and SAP Ariba from a Client Perspective
Managing Complex Services in SAP and SAP Ariba from a Client Perspective
 
Become Jythonic in FDMEE (KSCOPE15)
Become Jythonic in FDMEE (KSCOPE15)Become Jythonic in FDMEE (KSCOPE15)
Become Jythonic in FDMEE (KSCOPE15)
 
Planning Process Considerations
Planning Process ConsiderationsPlanning Process Considerations
Planning Process Considerations
 
How to play Golf on the first day of Go-Live for a JD Edwards EnterpriseOne ...
 How to play Golf on the first day of Go-Live for a JD Edwards EnterpriseOne ... How to play Golf on the first day of Go-Live for a JD Edwards EnterpriseOne ...
How to play Golf on the first day of Go-Live for a JD Edwards EnterpriseOne ...
 
Getting the Most Out of EPM: A deep dive into Account Reconciliation Manager
Getting the Most Out of EPM: A deep dive into Account Reconciliation ManagerGetting the Most Out of EPM: A deep dive into Account Reconciliation Manager
Getting the Most Out of EPM: A deep dive into Account Reconciliation Manager
 
Introduction to Oracle Hyperion Planning - New Features in 11.1.2.4
Introduction to Oracle Hyperion Planning - New Features in 11.1.2.4Introduction to Oracle Hyperion Planning - New Features in 11.1.2.4
Introduction to Oracle Hyperion Planning - New Features in 11.1.2.4
 
UGC NET Classroom coaching photographs
UGC NET Classroom coaching photographsUGC NET Classroom coaching photographs
UGC NET Classroom coaching photographs
 
Oracle Data Integration - Overview
Oracle Data Integration - OverviewOracle Data Integration - Overview
Oracle Data Integration - Overview
 
Unit 1 Overview of International Business
Unit 1 Overview of International BusinessUnit 1 Overview of International Business
Unit 1 Overview of International Business
 
Oracle Architecture
Oracle ArchitectureOracle Architecture
Oracle Architecture
 
Literal e texto-integro_de_contratos_colectivos_vigentes_oct_2015
Literal e texto-integro_de_contratos_colectivos_vigentes_oct_2015Literal e texto-integro_de_contratos_colectivos_vigentes_oct_2015
Literal e texto-integro_de_contratos_colectivos_vigentes_oct_2015
 

Similar to Leverage data analytics to reduce human space mission risks Presentation

Rapid Prototyping with Solr
Rapid Prototyping with SolrRapid Prototyping with Solr
Rapid Prototyping with Solr
Erik Hatcher
 
Michael_K_Bartosewcz_September_2015_resume
Michael_K_Bartosewcz_September_2015_resumeMichael_K_Bartosewcz_September_2015_resume
Michael_K_Bartosewcz_September_2015_resumeMike (Bart) Bartosewcz
 
Accelerating Science with Cloud Technologies in the ABoVE Science Cloud
Accelerating Science with Cloud Technologies in the ABoVE Science CloudAccelerating Science with Cloud Technologies in the ABoVE Science Cloud
Accelerating Science with Cloud Technologies in the ABoVE Science Cloud
Globus
 
OSFair2017 Workshop | EPOS: European Plate Observing System
OSFair2017 Workshop | EPOS: European Plate Observing SystemOSFair2017 Workshop | EPOS: European Plate Observing System
OSFair2017 Workshop | EPOS: European Plate Observing System
Open Science Fair
 
An Update on Arm HPC
An Update on Arm HPCAn Update on Arm HPC
An Update on Arm HPC
inside-BigData.com
 
Aerospace defensetechs
Aerospace  defensetechsAerospace  defensetechs
Aerospace defensetechsalancabe
 
GDPR compliance application architecture and implementation using Hadoop and ...
GDPR compliance application architecture and implementation using Hadoop and ...GDPR compliance application architecture and implementation using Hadoop and ...
GDPR compliance application architecture and implementation using Hadoop and ...
DataWorks Summit
 
Rolls-Royce Engineering Infographic
Rolls-Royce Engineering InfographicRolls-Royce Engineering Infographic
Rolls-Royce Engineering InfographicAli Mayar
 
Geo Analytics Canada Overview - May 2020
Geo Analytics Canada Overview - May 2020Geo Analytics Canada Overview - May 2020
Geo Analytics Canada Overview - May 2020
GEO Analytics Canada
 
Apache Metron in the Real World
Apache Metron in the Real WorldApache Metron in the Real World
Apache Metron in the Real World
DataWorks Summit
 
Deep Learning on Apache Spark at CERN’s Large Hadron Collider with Intel Tech...
Deep Learning on Apache Spark at CERN’s Large Hadron Collider with Intel Tech...Deep Learning on Apache Spark at CERN’s Large Hadron Collider with Intel Tech...
Deep Learning on Apache Spark at CERN’s Large Hadron Collider with Intel Tech...
Databricks
 
Apache Hadoop 3.0 What's new in YARN and MapReduce
Apache Hadoop 3.0 What's new in YARN and MapReduceApache Hadoop 3.0 What's new in YARN and MapReduce
Apache Hadoop 3.0 What's new in YARN and MapReduce
DataWorks Summit/Hadoop Summit
 
TERN Facility Portals - Stuart Phinn
TERN Facility Portals - Stuart PhinnTERN Facility Portals - Stuart Phinn
TERN Facility Portals - Stuart Phinn
TERN Australia
 
Embry Riddle Final
Embry Riddle FinalEmbry Riddle Final
Embry Riddle Final
jschrell
 
AusCover portal presentation
AusCover portal presentationAusCover portal presentation
AusCover portal presentationTERN Australia
 
Activeeon - Scale Beyond Limits
Activeeon - Scale Beyond LimitsActiveeon - Scale Beyond Limits
Activeeon - Scale Beyond Limits
Activeeon
 
A Data Lake and a Data Lab to Optimize Operations and Safety within a nuclear...
A Data Lake and a Data Lab to Optimize Operations and Safety within a nuclear...A Data Lake and a Data Lab to Optimize Operations and Safety within a nuclear...
A Data Lake and a Data Lab to Optimize Operations and Safety within a nuclear...
DataWorks Summit/Hadoop Summit
 
Overview and Status of HDF in NPOESS & NPP
Overview and Status of HDF in NPOESS & NPPOverview and Status of HDF in NPOESS & NPP
Overview and Status of HDF in NPOESS & NPP
The HDF-EOS Tools and Information Center
 
XSEDE14 SciGaP-Apache Airavata Tutorial
XSEDE14 SciGaP-Apache Airavata TutorialXSEDE14 SciGaP-Apache Airavata Tutorial
XSEDE14 SciGaP-Apache Airavata Tutorial
marpierc
 

Similar to Leverage data analytics to reduce human space mission risks Presentation (20)

Rapid Prototyping with Solr
Rapid Prototyping with SolrRapid Prototyping with Solr
Rapid Prototyping with Solr
 
Michael_K_Bartosewcz_September_2015_resume
Michael_K_Bartosewcz_September_2015_resumeMichael_K_Bartosewcz_September_2015_resume
Michael_K_Bartosewcz_September_2015_resume
 
Accelerating Science with Cloud Technologies in the ABoVE Science Cloud
Accelerating Science with Cloud Technologies in the ABoVE Science CloudAccelerating Science with Cloud Technologies in the ABoVE Science Cloud
Accelerating Science with Cloud Technologies in the ABoVE Science Cloud
 
OSFair2017 Workshop | EPOS: European Plate Observing System
OSFair2017 Workshop | EPOS: European Plate Observing SystemOSFair2017 Workshop | EPOS: European Plate Observing System
OSFair2017 Workshop | EPOS: European Plate Observing System
 
An Update on Arm HPC
An Update on Arm HPCAn Update on Arm HPC
An Update on Arm HPC
 
Aerospace defensetechs
Aerospace  defensetechsAerospace  defensetechs
Aerospace defensetechs
 
GDPR compliance application architecture and implementation using Hadoop and ...
GDPR compliance application architecture and implementation using Hadoop and ...GDPR compliance application architecture and implementation using Hadoop and ...
GDPR compliance application architecture and implementation using Hadoop and ...
 
Sword Crig 2007 12 06
Sword Crig 2007 12 06Sword Crig 2007 12 06
Sword Crig 2007 12 06
 
Rolls-Royce Engineering Infographic
Rolls-Royce Engineering InfographicRolls-Royce Engineering Infographic
Rolls-Royce Engineering Infographic
 
Geo Analytics Canada Overview - May 2020
Geo Analytics Canada Overview - May 2020Geo Analytics Canada Overview - May 2020
Geo Analytics Canada Overview - May 2020
 
Apache Metron in the Real World
Apache Metron in the Real WorldApache Metron in the Real World
Apache Metron in the Real World
 
Deep Learning on Apache Spark at CERN’s Large Hadron Collider with Intel Tech...
Deep Learning on Apache Spark at CERN’s Large Hadron Collider with Intel Tech...Deep Learning on Apache Spark at CERN’s Large Hadron Collider with Intel Tech...
Deep Learning on Apache Spark at CERN’s Large Hadron Collider with Intel Tech...
 
Apache Hadoop 3.0 What's new in YARN and MapReduce
Apache Hadoop 3.0 What's new in YARN and MapReduceApache Hadoop 3.0 What's new in YARN and MapReduce
Apache Hadoop 3.0 What's new in YARN and MapReduce
 
TERN Facility Portals - Stuart Phinn
TERN Facility Portals - Stuart PhinnTERN Facility Portals - Stuart Phinn
TERN Facility Portals - Stuart Phinn
 
Embry Riddle Final
Embry Riddle FinalEmbry Riddle Final
Embry Riddle Final
 
AusCover portal presentation
AusCover portal presentationAusCover portal presentation
AusCover portal presentation
 
Activeeon - Scale Beyond Limits
Activeeon - Scale Beyond LimitsActiveeon - Scale Beyond Limits
Activeeon - Scale Beyond Limits
 
A Data Lake and a Data Lab to Optimize Operations and Safety within a nuclear...
A Data Lake and a Data Lab to Optimize Operations and Safety within a nuclear...A Data Lake and a Data Lab to Optimize Operations and Safety within a nuclear...
A Data Lake and a Data Lab to Optimize Operations and Safety within a nuclear...
 
Overview and Status of HDF in NPOESS & NPP
Overview and Status of HDF in NPOESS & NPPOverview and Status of HDF in NPOESS & NPP
Overview and Status of HDF in NPOESS & NPP
 
XSEDE14 SciGaP-Apache Airavata Tutorial
XSEDE14 SciGaP-Apache Airavata TutorialXSEDE14 SciGaP-Apache Airavata Tutorial
XSEDE14 SciGaP-Apache Airavata Tutorial
 

Leverage data analytics to reduce human space mission risks Presentation

  • 1. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation Haden Land haden.land@lmco.com Jason Loveland jason.c.loveland@lmco.com
  • 2. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Orion Program Overview  High Data Rate Ingest Use Case  MACH-5 INSIGHT℠ Platform  SFDU and Protobuf  HBase bulk loading using Spark Streaming  TRACE Visualization
  • 3. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation Orion Multi-Purpose Crew Vehicle, NASA’s first spacecraft designed for long-duration, human-rated deep space exploration. Orion will transport humans to interplanetary destinations beyond low Earth orbit, such as asteroids, the moon and eventually Mars, and return them safely back to Earth.
  • 4. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Carry out a robust human and robotic exploration program that is both sustainable and affordable  Extend human presence across the solar system to our planets, asteroids and other destinations  Develop the innovative technology, knowledge, and infrastructure needed to support more challenging human space exploration missions  Promote international and commercial participation in exploration to further advance U.S. scientific, security and economic interests
  • 5. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Uncrewed Orion launched from Cape Canaveral Air Force Base on a Delta IV Heavy rocket  Orbited the Earth twice, traveling approximately 3,600 miles above the Earth’s surface • 15 times farther than the International Space Station. • Generated more than 80% of the return velocity experienced during a reentry from the moon, which allows engineers to model expected reentries from future missions in deep space.  Orion travelled twice through the Van Allen Radiation Belt, a dense radiation field of highly energetic charged particles surrounding Earth • This exposure will help the team understand and mitigate radiation exposure, and to develop protective solutions before the first crewed mission.  At the conclusion of EFT-1, Orion reentered Earth’s atmosphere at 20,000 mph and enduring temperatures of up to 4,000° Fahrenheit  After the parachutes deployed, the spacecraft slowed to 20 mph splashing down in the Pacific Ocean, near the Baja Peninsula.
  • 6. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
  • 7. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation • How much telemetry is there, anyway? Is it “big data”? • Orion is transitioning to Big Data • LOTS! PB+. On board, three 1 Gb networks are kept full • Volume, velocity, variety • What is Spacecraft Telemetry? • Sensor measurements to monitor everything • Microsecond Timestamps • 12 discrete time sources • More complex time series • Human Space Flight requires fault tolerance • What produces telemetry? • For Orion, the telemetry is produced in a variety of simulation and test environments which includes at least 7 differ labs across the US. • How many items are there to measure in Orion? • The XTCE representing the flight software telemetry has over 3,000,000 elements, of which over 300,000 are related to currently taken measurements. There are many more when considering the ground system and other subsystems. • How is telemetry analyzed today? • Traditionally, the responsible engineer watches a handful of the telemetry indicators on the real time monitors as a test progresses, or in review. Certain behavior and reporting will result in further analysis on a few other measurands. • Most values of most telemetry measurements are ignored if they are not out of limits • Specific studies are done on some measurements in historical context, usually after detection of anomalous behavior, to determine if the behavior has been observed in the past. • Automation has been applied to the problems of automatically collecting data, automatically running scripted tests, and automatically detecting out-of- limit values
  • 8. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
  • 9. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation Orion’s next mission (EM-1) in 2018  2 weeks instead of 4 hours  4 times as many computers  Twice as many instruments  Subsystems that support Human Flight! EM-1 will be orders of magnitude bigger and more
  • 10. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Technologists/Platform Developers • Networking • Disk storage strategies • Modern languages and techniques: Scala, R • Big Data tools: Hadoop ecosystem, NoSQL databases, Stream and Batch Processing  Analytics /Data Scientist(Ponderers) • Tools of the trade • Visualization & Communication • Data Science  Statistics  Design of experiments  Statistical tools: SAS, SPSS, R, MATLAB  Modeling  Data Mining  Linear algebra  Data cleansing, sampling, warehousing  Subject Matter Experts • System and Subsystem Experts • Program Tools Experts
  • 11. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation Horizontally scalable, fault tolerant prototype for processing High Data Rate telemetry from Orion Test Rigs including ingest, processing, transform, storage, advanced analytics  Satisfy Data Rate: ~300MB/s (97MB/s X 3 planes)  Provide common data wire protocol that is:  Flexible, expressive and compact  Programming language agnostic  Provide system to scale telemetry ingest from Test Rigs to Mach-5 within technical performance metrics  Provide solution that is continuously measurable and monitorable  Provide Advanced Analytics to identify anomalies on ingest
  • 12. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
  • 13. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Enables Orion to teach us lessons about it’s own experience leveraging data to tell the story  Reduces time to develop and test Orion by identifying failures early!  Uses data points from one mission to help predict if mission critical components will succeed or fail in specific future mission circumstances  Permits queries to be retrieved across historical batch and live streaming views within seconds  Streams historical data back as if it were live  Overlays data across tests and missions  Stores and processes much larger data sets than any other space craft in history Scales to meet the needs of future missions to keep the conversation all the way to Mars
  • 14. Data Reader/Simulator Ingest Packet Measurands (GPBs) Kafka Message Bus Packet Measurands (GPBs) Deduplicat ion (Spark) HBase Writer (Spark) mach5- sample ObjSplitter + Decom (GDS) C++ client Reads Packets and Decommutates Tlm Data Packet Measurands GPB File (represents a Packet(s) and contains decommutated measurands) Header Metadata apid:seqctr:time: value1 … .. apid:seqctr:time: valueN mach5- sample (Spark) Packet Measurands (GPBs) Lockheed Martin Proprietary Information Storage Analytics HDFS HFiles (HBase-RDD) Mach-5 Data Ingest for Orion HBase Web/UITomcat Glassfish Etc. Trace FOSS widgets Aggregatio n (Spark) Alerting (Spark) Limit Checking (Spark)
  • 15. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation • Leverage space standards and wrapping modern data serialization technologies • SFDU (Standard Formatted Data Unit) • CCSDS Standard • Consultative Committee for Space Data Systems – Founded 1982 • Delimiting data and providing an internal index • Transport and archival format • Provides ease of integration with space systems • Should be usable 25+ years from now • SFDU wraps serialized data, in our case Protobuf CCSD Z 001L LMCO U 009L V V LMCO S 010 LMCO I 010 L L V V Header Body
  • 16. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Protocol Buffers • Compact, compressible • Many language bindings: Java, C++, Python, etc… • Self describing schema, auto-generates code • Compared to other Serialization libraries (i.e. Avro), allows for greater data type expression.  Ex: Can natively support unsigned Long and Unsigned Integer on C++ side  Allows for cleaner generated code bindings • Widely used within the Hadoop ecosystem • Widely used within Lockheed Martin telemetry processing
  • 17. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Enables Orion to scale data architecture linearly on FOSS  Allows for multiple test rigs to simultaneously write data to be ingested at high data rates  Provides ability for analytics processing to independently pull and push data  Provides replay capabilities for guaranteed, exactly once processing semantics per partition  Provides fault tolerance and replication
  • 18. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Enables Orion to scale data architecture linearly on FOSS  Real-time Analytics reduce time to develop and test the Orion by identifying anomalies early!  Allows for analytics to be run in batch and streaming mode  Analytics libraries that scale  Integration with Kafka to independently pull and push data  Runs on YARN to reduce complexity of managing resources  Provides guaranteed, exactly once processing semantics per partition using Kafka Direct Connect  Provides fault tolerance and check pointing  Integration with HBase (SparkOnHBase)
  • 19. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation Pull Data off of Kafka and Transform from SFDU to Protobuf HBase Column Format •Data bucketed to be aligned to appropriate Region. (HBase table is pre-split). Handled by Spark on Yarn Sort Rows and Partitioning by HBase Regions •Sort happens locally •HFiles are guaranteed to contain all rows for a specific region •This avoids CPU intensive work by HBase Construct HFiles Write to HDFS •HFiles are written to HDFS •Writing HFiles is much faster than HBase puts Bulk Load into HBase •Bulk Load is called on each HFile
  • 20. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Data arrives in random order and HBase automatically sorts data by time  Enables Orion to scale data architecture linearly on FOSS  Provides random access to all telemetry with low latency access  Provides data store to perform batch processing using Mapreduce and Spark  Custom filters
  • 21. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Single-Column Schema • Single Colum is packed with every piece of metadata for a row. Numerous serialization formats can be used: json, protobuf, avro, etc…  Benefits • Compact, greatly decreases row footprint • Higher write throughput. Under the hood HBase performs a Put for each column of a row. We go from ‘N’ puts in a multi column schema to 1 • Caching for queries  Detractors • Increased overhead to update the column • If your column schema has the potential to change, could end up with multiple versions of the serialized column value • Slower reads because requires that all of the data be returned and deserialized • Requires a custom HBase filter to filter on individual members of the column value • Integration with 3rd Party BI tools and the rest of the Hadoop ecosystem
  • 22. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation Detect abnormal activity during a test and notify operators early as possible. (Tests can run for weeks!)  Provide limit checking on individual measurements  Combine measurements to compute higher level understanding (i.e. momentum calculation)  Aggregation and statistical summaries  Requirements validation
  • 23. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation
  • 24. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation  Analytics!  We need the help of the community to develop effective data platforms and analytics that can scale with Orion missions.
  • 25. © 2015 Lockheed Martin Corporation LOCKHEED MARTIN and LOCKHEED trademarks used throughout are registered trademarks in the U.S. Patent and Trademark Office owned by Lockheed Martin Corporation

Editor's Notes

  1. Tools: SAS, SPSS, R, MATLAB, Python+SciPy, SQL, natural language processing, Baysian models, Petri nets Visualization and communication: : Tableau, D3 or other JS, Cesium
  2. March 24th Delivery Ingest Spark Job Testing and hardening Deduplication Spark Job Testing and hardening Aggregation Spark Job Testing and hardening Limit Checking Spark Job Testing and hardening Alerting Spark Job Prototyped TRACE integration and installation