This document discusses dashDB, a cloud data warehouse service from IBM that provides fully managed analytics capabilities including geospatial support. Key features highlighted include: - In-database analytics built on DB2 BLU and Netezza technologies for fast performance. - Integrated R analytics and predictive modeling capabilities. - Support for geospatial data and analytics using Open Geospatial Consortium standards. - Plans available as fully managed cloud service, on-premises, or hybrid architectures.

1. dashDBthe GIS professional’s bridge to mainstream IT systems
John Park
Product Manager
IBM
jjpark@ca.ibm.com

2. Hi IBM … what have
you been up to lately ?

3. Why am I here, and what am I going
to talk about ?
dashDB and ArcGIS support
Show you stuff

4. ©2015 IBM Corporation
Supporting Applications at Web Scale
Fast, fully managed, cloud data warehouse service
Integrated analytics
R predictive modeling
Geo, Open and Twitter Data Support
Built-in RStudio IDE
Built-in performance with in-memory
technology
Predictive modeling built into the database
(linear regression, k-means clustering, Esri
compatible, and more)
Works with an ecosystem of apps and tools
Integrated security and maintenance
IBM dashDB

5. ©2015 IBM Corporation
For apps that need:
• Elastic scalability
• High availability
• Data model flexibility
• Data mobility
• Text search
• Geospatial
Available as:
• Fully managed DBaaS
• On-premises private cloud
• Hybrid architecture
BLU
Acceleration
Netezza
In-Database
Analytics
Cloudant
NoSQL
Integration
dashDB MPP
Fully-managed data warehouse on cloud
 DB2 BLU columnar technology +
Netezza in-database analytics
 BLU in-memory processing, data skipping, actionable
compression, parallel vector processing, , “Load & Go”
administration
 Netezza predictive analytic algorithms,
fully integrated RStudio & R language
 Oracle compatibility
 Massively Parallel Processing (MPP)
dashDB
In-database analytics capabilities for best performance atop a fully-managed warehouse

6. ©2015 IBM Corporation
BLU is better

7. ©2015 IBM Corporation
Massively Parallel Platform
IBM Netezza
In-Database
Analytics
Transformations
Geospatial
Predictive
Statistics
Data Mining
More Tools
In-Database
Analytics
SAS
R
Fuzzy Logix
Zementis
IBM SPSS
BI Tools
Visualization
Tools
Software
Development
Kit
User-Defined
Extensions
(UDF, UDA,
UDTF, UDAP)
Language
Support
(Map/Reduce,
Java, Python,
Lua, Perl,
C, C++,
Fortran)
Custom Stored
Procedures
(NZPLSQL)
BigInsights
Fuzzy Logix
Streams
Apache Hadoop
Mathematics
Time Series
IBM Netezza Analytics at a Glance

8. ©2015 IBM Corporation
dashDB Key Use Cases
• Extend on-premises data warehouse
environments to the cloud
• Flexible, cost-effective growth
• Hybrid cloud models support ground to cloud
Extend / Modernize
• Easy synchronization of JSON to structured data
• Allows analytics via standard BI tools
• In-database predictive algorithms allow greater
insight for Cloudant users than ever before
JSON Analytics
• Robust predictive analytic algorithms
• Integrated with R
• Watson Analytics ready
• Analytics ecosystem with partners
In-Database
Analytics
• Data warehousing and analytics in the cloud
• Cloud agility and flexibility
• Analytics for cloud data, data marts, and
development & test environments
Data Warehouse &
Analytics Service

9. ©2015 IBM Corporation9

10. ©2015 IBM Corporation
Use Case: Car Manufacturer - Architecture Details
Vehicle Data
Other Data Sources
IBM Cloudant
GeoJSON data to relational
Cloudant Schema Discovery Protocol
IBM dashDB
IBM Dataworks

11. Enterprise MPP
• Dedicated, single tenant environment
• Bare metal
• 3+ node clusters
• 24 cores per node
• 256 GB memory per node
• SSD storage (for about 4 TB of
preload data per node)
• $5410 / month (USD) per node
Entry
• Shared, multi-tenant
environment
• 20 GB SAN storage
capacity
• Freemium: < 1 GB of
raw data is free
• $50 / month (USD)
flat rate
Enterprise - 1TB
• Dedicated, single
tenant environment
• Virtual environment
• 16 cores
• 64 GB memory
• SAN storage (for about
1 TB of preload data)
• $1170 / month (USD)
Enterprise - 4
TB
• Dedicated, single
tenant environment
• Bare metal
• 32 cores
• 256 GB memory
• SAN storage (for about
4 TB of preload data)
• $4700 / month (USD)
Enterprise - 12
TB
• Dedicated, single
tenant environment
• Bare metal
• 32 cores
• 256 GB memory
• SAN storage (for about
12 TB of preload data)
• $7370 / month (USD)
dashDB Plans Suit a Variety of Needs

12. Geospatial Analytics In dashDB
Implements OGC SFS & ISO SQL/MM part 3 standards for spatial
See http://www.iso.org/iso/catalogue_detail.htm?csnumber=38651
Spatial data type ST_GEOMETRY (hierarchy)
Enables spatial operations (e.g. joins) in database through spatial
operators available as user defined functions
Dedicated support in ESRI tools starting V 10.3
dashDB - R support through extension to ibmdbR package

13. Spatial Functions & Predicates in dashDB
SELECT a.name, a.type
FROM highways a, floodzones b
WHERE ST_Intersects(a.location,b.location) = 1
AND b.last_flood > 1950
SELECT a.road_id, a.time, i.id,
ST_Distance(a.loc, i.loc,’METER’) as distance
FROM accidents a, intersections i
WHERE ST_Distance(a.loc,i.loc,’METER’) < 10000
AND a.weather = ‘RAIN’
- accidents near intersections
- highways in flood zones
ST_Distance(g1,g2)
?
ST_Intersects(g1,g2)
?

14. Spatial Constructor Functions
ST_Point(x, y, srs_id) – create point at this location
ST_Point(‘POINT (-121.5, 37.2)’, 1)
ST_Linestring(‘LINESTRING (-121.5 37.2,-121.7 37.1)’,1)
ST_Polygon(CAST (? AS CLOB(1M)),1)
For host variable containing well-known text, well-known binary, or shape representation

15. Spatial Predicates – WHERE Clause
ST_Distance(geom1, geom2) < distance_constant or var
ST_Contains(geom1, geom2) = 1
ST_Within(geom1,geom2) = 1
EnvelopesIntersect(geom1, geom2) = 1
EnvelopesIntersect(geom1, x1, y1, x2, y2, srs_id) = 1
ST_Area(geom) < some_value

16. Spatial Functions that Create New Spatial Values
ST_Buffer(geom, distance)
ST_Centroid(geom)
ST_Intersection(geom1, geom2)

17. And Many More …
ST_Area
ST_AsBinary
ST_AsText
ST_Boundary
ST_Buffer
ST_Centroid
ST_Contains
ST_ConvexHull
ST_CoordDim
ST_Crosses
ST_Difference
ST_Dimension
ST_Disjoint
ST_Distance
ST_Endpoint
ST_Envelope
ST_Equals
ST_ExteriorRing
ST_GeomFromWKB
ST_GeometryFromTex
t
ST_GeometryN
ST_GeometryType
ST_InteriorRingN
ST_Intersection
ST_Intersects
ST_IsClosed
ST_IsEmpty
ST_IsRing
ST_IsSimple
ST_IsValid
ST_Length
ST_LineFromText
ST_LineFromWKB
ST_MLineFromText
ST_MLineFromWKB
ST_MPointFromText
ST_MPointFromWKB
ST_MPolyFromText
ST_MPolyFromWKB
ST_NumGeometries
ST_NumInteriorRing
ST_NumPoints
ST_OrderingEquals
ST_Overlaps
ST_Perimeter
ST_Point
ST_PointFromText
ST_PointFromWKB
ST_PointN
ST_PointOnSurface
ST_PolyFromText
ST_PolyFromWKB
ST_Polygon
ST_Relate
ST_SRID
ST_StartPoint
ST_SymmetricDiff
ST_Touches
ST_Transform
ST_Union
ST_WKBToSQL
ST_WKTToSQL
ST_Within
ST_X
ST_Y
And more…
Simplified
Constructors from
x,y
WKT
WKB
GML
shape
Linear referencing
Spatial aggregation
ST_AsGML
ST_AsShape

18. Harness the Full Power of SQL
Outer join
Common table expressions
Recursive queries, sub-queries
Aggregate functions
Order by, group by, having clauses
OLAP, XML, and more ...
WITH sdStores AS (SELECT * FROM stores
WHERE st_within(location, :sandiego) = 1)
SELECT s.id, s.name, AVG(h.income) FROM houseHolds h, sdStores s
WHERE st_intersects(s.zone, h.location) = 1
GROUP BY s.id, s.name
ORDER BY s.name
Example problem: Determine the average household income for the sales zone of
each store in the San Diego area.

19. GeoData & dashDB
{GeoJSON}
WKT((),())
WKB
GML

20. Let me show you something

21. Merci Beaucoup !
dashDB.com
ibm.biz/openfordata