
Sivabalan Narayanan
Education
- B.E, PSG College of Technology(Affiliated to Anna Univ), Coimbatore, India
Bio
A Distributed systems enthusiasist.
Employment:
• Ayasdi Palo Alto, CA, US
Intern – Software Engineering Jun 2013 – Sep 2013
• Oracle Bangalore, India
Member Technical Staff Sep 2009 – Aug 2012
Beehive Collaboration Suite (Server Technology) & Fusion Service CRM Application
Languages: Java, C++, C
Framework and Database: Rabbitmq, Protocol Buffers, Memcached, HBase, Hadoop, Map-Reduce, Celery, Redis
Projects
• Message Futures : Multi-Datacenter Replication
• TOPCOP : TwO Phase Commit Over Paxos (Spanner)
• C-SyTS : Cloud based multi-tasking System for Task Scheduling
• GEMS : Gossip-Enabled Monitoring Service for Heterogeneous Distributed Systems
• K-NN based Music Recommendation System
Honors and Awards
- Honored the “Best Outgoing Student Award” and a Gold Medal for the batch 2005-2009
- Final Year project, Linguistic Engine was chosen as the “Best Project”
- Oracle Certified Java Programmer
- Presented the “Star Sporstsman of the Year Award” for appearance in National Sepak Takraw Championship
Research
• Message Futures : Multi-Datacenter Replication - UCSB
- Developed a Multi-Datacenter Replication system which had a strongly consistent concurrency control manager with low commit latency to ensure
mutual consistency of replicas across datacenters.
- Matrix clocks were used to maintain synchronization among replicas.
- Website: http://www.cs.ucsb.edu/~sivabalan/cs274
• TOPCOP : TwO Phase Commit Over Paxos (Spanner) - UCSB
- Built a distributed storage system to carry out transactions in a consistent manner.
- To strike a balance between consistency and availability in a distributed environment, Two Phase Commit protocol was merged with Paxos consensus algorithm.
- Website: http://www.cs.ucsb.edu/~sivabalan/cs271
• C-SyTS : Cloud based multi-tasking System for Task Scheduling - UCSB
- Designed a feature rich pluggable task scheduler that could be deployed in a Cloud based Multitasking System.
- Dynamic scale up or scale down of nodes based on resource usage (Auto-scaling).
- Website: http://cs.ucsb.edu/~sivabalan/cs263
• GEMS : Gossip-Enabled Monitoring Service for Heterogeneous Distributed Systems - UCSB
- GEMS is a highly responsive and scalable resource monitoring service, to monitor the liveness of nodes in a heterogeneous distributed systems and
to achieve consensus in a distributed environment.
- Website: http://cs.ucsb.edu/~sivabalan/gems
• K-NN based Music Recommendation System – UCSB
- Developed a music recommendation system which recommends songs for a user based on user’s behavior. K nearest neighborhood approach is followed
as a collaborative filtering algorithm to find nearest neighbors for the item in context.
- Website: www.cs.ucsb.edu/~sivabalan/cs240a