2. First and foremost…
We are a software company
We are not systems integrators…
We are not a consultancy…
We are not a bespoke development center…
Who are we?
3. First and foremost…
We are a software company
Who are we?
We are innovators
We are humble and hungry
We do have a point of view
4. We build the best software solutions we can
Simple, but not simpler
Usability comes first
Resilient
Continuous improvement
What do we do?
5. Stack for Streaming Data…
SYSLOG
SNMP
ETW
Sources
.
.
Agent A
Agent B
Agent N
Capture
.
.
A Topic
B Topic
N Topic
Distribution
.
.
Topology
A
Topology
B
Topology
N
Real Time
.
.
Hive
Elastic
HBase
Storage
Analytical
Tools
UI
Web
Access
Access
7. Grumpy cat questions?
How do we…
… add security? … ensure best performance?
… add auditability? … work with data abstractions?
… monitor it? … add new algorithms?Σ
8. Maybe it is time to…
…Think again
Simple, but not simpler
Usability first
Resilient
Continuous improvement
10. I would like…
Streaming system
Strong data centric solution
Grumpy cat shopping list
Rich type system
Analytical results are streams
Multiple data formats
Data validation
Immutable
11. I would like…
Streaming system
Strong data centric solution
One cluster
Grumpy cat shopping list
Decentralization
Symmetry
Incremental scalability
Heterogeneous
Partition & replication
Administrator friendly
12. I would like…
Streaming system
Strong data centric solution
One cluster
Open
Grumpy cat shopping list
APIs everywhere
REST architecture
13. I would like…
Streaming system
Strong data centric solution
One cluster
Open
ML friendly
Grumpy cat shopping list
Write algorithms only once
Compatible with tools like R, Matlab, Python
14. I would like…
Streaming system
Strong data centric solution
One cluster
Open
ML friendly
Secure & auditable
Grumpy cat shopping list
Perfect forward secrecy
Data encrypted on disk
Clean monitoring
17. Before you attempt to replicate our craziness…
This is hard. Seriously hard.
Amazing team
Still… If you are crazy enough…
Word of advice
18. Valo …
… is a dynamo inspired cluster
… incorporates built-in repositories
… exposes a rich API and data set
… polyglot
… unifies historical and stream processing
So, what is Valo?
20. Future
Is Valo finished?
No! Continuous improvement, remember?
Focus on ML algorithms
Add an J/ODBC driver
Python / R integration
Formally publish all the SDKs
Data acquisition