Josep Lluis Larriba-Pey – Graph databases go mobile, Sparksee 5 mobile use cases
The use of graph databases is becoming more and more popular. Sparksee is a clear example of a High Performance graph database that allows for high compression and small software footprint, allowing for very compact and efficient solutions in the business world. Sparsity Technologies has recently released Sparksee 5 mobile for iOS, Android and BB10, allowing for high performance mobile applications to be boosted with high performance analytics. Sparksee is a research based software with a considerable number of papers published on it, showing the importance of research in high end technologies.In this talk, we will present Sparksee 5 mobile and explain a few use cases in the area of analytics for Social and Open Data where the use of graphs boosts job search, private recommendation, community search and personal tourist route planning.
Call Girls In Doddaballapur Road ☎ 7737669865 🥵 Book Your One night Stand
Josep Lluis Larriba-Pey – Graph databases go mobile, Sparksee 5 mobile use cases - NoSQL matters Barcelona 2014
1. Sparksee Graph Database
Graph Databases go mobile
Sparksee 5.1 use cases
November 2014 Josep Lluis Larriba-Pey
º
*SparsityTechnologies — Powering Extreme Data sparsity–technologies.com
2. A tip and two questions º Sparksee Graph Database
A tip and two questions
The tip
• Barcelona, the first NoSQL city back in the XII century (ACA XIV, RN XIII,
ABCN XIII)
• Finances, Taula de Canvi, first public Bank, with a historic archive, banker
• Casa de Convalescència cost 18K€
First question
• I need you to help me, do you want to appear in Sparsity’s Twitter?
• Do you want me to keep talking about History or shift to NoSQL?
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
3. Sparksee º Sparksee Graph Database
Graphs everywhere!
Social
E-mail
Maps
Apps
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
4. Large vs Small scale graph analytics º Sparksee Graph Database
Large vs small scale graph analytics
Large
• Obtained from the community use of resources
• Provide a global view of users for companies
• Require the use of large servers
• Can be used for BI or Analytics
Small
• Obtained from the local use of resources
• Smaller scale view of “my” world
• Can be managed in smaller scale devices
• Can be used for my own needs
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
5. Roadmap º Sparksee Graph Database
Roadmap for this talk
1. Sparksee 5.1
2. Sparksee 5.1 mobile
3. Use cases for mobile devices
4. Social Network Analytics at Sparsity
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
6. 1. Sparksee 5.1 º Sparksee Graph Database
1. Sparksee 5.1
IS a high-performance and out-of-core graph database management
system
FOR large scale labeled and attributed multigraphs
Labeled: Nodes and edges belong to types
Attributed: Nodes and edges may have attributes
Multigraph: Several edges between nodes
BASED ON vertical partitioning and collections of objects identifiers
stored as bitmaps
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
7. 1. Sparksee 5.1 º Sparksee Graph Database
Sparksee Characteristics
Graph split into small structures
Move to main memory just significant parts (caching)
Object identifiers (oids) instead of complex objects
Reduce memory requirements
Specific structures to improve traversals
Index the edges and the neighbors of each node
Attribute indices
Improve queries based on value filters
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
8. 1. Sparksee 5.1 º Sparksee Graph Database
Sparksee Architecture
— Database core (C++)
— Low level C++ layer: It allows a direct
interaction with applications coded in
C++
— Interface layer (SWIG*): Wrapper
generator for those APIs that are not
C++.
— APIs for Python, Java and .Net on top
of our C++ API
— Mobile APIs for iOS, Android and
BB10.
.Net
App
Native APIs
SWIG
SparkseeC++ - Graph Algorithms
GDB
GRAPH
DATA
JAVA
App
BUFFERPOOL
PLATFORM
SPKSEECORE
BB10
App
Python
App
C++
App
Objective
C
App
Android
App
*SWIG = Simplified Wrapper and Interface Generator.
Open source tool used to connect programs/libraries written in
C/C++ with other languages.
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
9. 2. Sparksee 5.1 Mobile º Sparksee Graph Database
2. Sparksee 5.1 Mobile
FIRST EVER
Full fledged Graph Database in your mobile device
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
10. 2. Sparksee 5.1 Mobile º Sparksee Graph Database
Characteristics
• Small code footprint (less than 75Kbytes).
• ACID properties, fully transactional.
• Provides APIs for most platforms, i.e. Android, iOS and BB.
• Fast execution of complex queries (low battery cost).
Advantages
• Sparksee empowers your application at a minimal memory cost.
• Connection may be interrupted amid a transaction and you may roll it back.
• You may have your device disconnected and still be able to analyze data.
• Analytics power in your mobile device.
• For App vendors, someone else spends resources and I get the result of the analysis
• Community search: fastest and best quality (WWW’14, TETRACOM)
Roadmap
• Provide sync with your server database through Sparksee.
• Provide a set of APIs that solve most of your problems: Integration, Social Analytical API,
BI API, etc.
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
11. 3. Use cases formobile devices º Sparksee Graph Database
3. Use cases for mobile devices
Mobile device assets
• Data integration (e-mail, contacts, Social Networks, Apps metadata)
Added value with my own data
• My Social Network analytics
• My CRM
Reliability/availability of the communication
• Medical environment
• Travel
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
12. 3. Use cases formobile devices/Mobile device assets º Sparksee Graph Database
Data integration
• Data in a mobile device:
Social
Apps Contacts
• Closely related data
• Managing all your data from a single analytical full fledged database
• It provides an integrated view of all the content in your virtual office
• Easy to query for patterns, relationships, and other complex
algorithms
E-mail
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
13. 3. Use cases formobile devices/Added value with my own dataº Sparksee Graph Database
My Social Network analytics
• Messaging:
• Sending the message to the adequate people…
• What are the SNs that will secure a better coverage?
• Who can be more interested and has a better impact in the SNs?
• Who is knowledgeable about the topic, so that I can ask for opinion?
• Recommendation
• Collaborative filtering with my own data, let the graph know aboutme and
my friends!
• How can I group my acquaintances to go out tonight? For instance:
• People who like the same type of music and get on well among them…
• People who usually go to close by places…
• People who always get involved in heated discussions…
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
14. 3. Use cases for formobile mobile devices/devices Added value with my own data
º Sparksee Graph Database
My CRM
• I have access to my professional social networks via API
• I want to sell a product…
• … want to invite people to an event
• I have to find common friends who are customers
• From my own Social Networks’ contacts
• I want to manage my contacts to know how they evolve in their tastes
• How are they related?
• What patterns can I infer from their evolution?
• From my own Social Networks
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
15. 3. Use cases formobile devices º Sparksee Graph Database
Reliability/availability of communications
• Map management
• Geo positioning management
• Movement detection
• Other information:
• POIs
• Medically oriented patterns
• Learning your tastes
Maps
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
16. 3. Use cases formobile devices/Reliability-availability of commºunications Sparksee Graph Database
Medical environment
• Specific illnesses require constant monitoring
• Detecting the gradient of the movement
• Predicting dangerous patterns
• For example:
• I am diabetic1 and I am hiking in the mountains
• No data connection, maps in device with height water-marks
• My application detects a pattern, predicts danger and warns me:
• Dangerous walking gradient predicted though map
• Before heartbeat increase starts, the system warns me and
gives me advice
165% of diabetics die of heart disease and stroke
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
17. 3. Use cases formobile devices º Sparksee Graph Database
Travel
• Proposing specific itineraries
• My local maps, POIs and coupons are
downloaded when wi-fi available
• Off-line re-computation of itineraries:
• I move away from the itinerary proposed
• MyApp computes a new route based on local map
• Commercial recommendations based on your Social off-line “taste” and learning
• Recommend restaurants
• Possibility to redeem coupons
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
18. 4. Social Network Analytics at Sparsity º Sparksee Graph Database
4. Social Network Analytics at Sparsity
Will provide a SaaS SNA platform
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com
19. 4. Social Network Analytics at Sparsity º Sparksee Graph Database
Cloud platform
• Connectors
• Twitter, RSS, Parsers
• You’ll be able to create and include other connectors
• Technologies for analytics
• Community detection based on fastest and most accurate techs.
• Role detection, Entity recognition, Sentiment analysis
• Open SaaS API
• Synchronous and assynchronous connection for services on data
*SparsityTechnologies ——Powering Extreme Data º sparsity––technologies.com