Second IMC Pune Meetup

Hello again IMC Enthusiasts,

Let's get together in the new year for our second In Memory Computing meet-up in Pune.

[Note: This is a FREE meetup.]

Lining-up some great speakers for the event - following is the developing agenda:

AGENDA:

6:00 PM - 6:30 PM : Welcome & Networking

6:30 PM - 7:00 PM : Non Volatile Memory & Implications on Data Infrastructure - Shankar Hundekar/ Robert Geiger (Ampool)

7:00 PM - 7:30 PM : Snacks & Networking

7:30 PM - 8:00 PM : Apex & Geode: In-memory Streaming on Hadoop - Sandeep Deshmukh (DataTorrent) & Ashish Tadose (Ampool)

8:00 PM - 8:30 PM : Disorderly programming: Avoiding cost of enforcing time order - Shripad Agashe (Thoughtworks)

8:30 PM - 9:30 PM : Networking/ Wrap-up

ABSTRACTS & BIOS:

Talk 1: Non Volatile Memory & Implications on Data Infrastructure

In this talk, we describe the latest trends and advancements in storage-class memory (a.k.a. Non-Volatile Memory) and its impact on data infrastructure as it exists today. We also describe how Ampool is building a product that allows Big Data analysis solutions to work together with a smart storage class memory layer to allow fast & complex end to end analytical pipelines, thus lowering the time-to-insights significantly.

BIOs:

Shankar Hundekar

Before joining Ampool India Pvt. Ltd. as a Senior Software Engineer, Shankar worked with VMWare/Pivotal as Member of technical staff on different products like GemFire, GemFireXD. He worked on various features/improvements on Native Clients for GemFire and also worked on building ODBC client driver GemFireXD. Prior to VMware he worked in Symantec on Enterprise Vault, Discovery Accelerator, Compliance Accelerator and Clearwell ediscovery. Before joining Symantec he worked in CDAC R&D in GIST group on text to speech syllable splitter and ISM product. Shankar has bachelor's degree in Computer Science from Government College Of Engineering Aurangabad. He has also done PG Diploma in Advanced Computing from CDAC Pune.

Robert Geiger

Before joining Ampool, Robert was an Architect at B2B deep learning start-up, and prior to that as an Architect and team lead at Pivotal, working on security and analytics as a service for the Hadoop ecosystem. He possesses broad knowledge of distributed systems, databases, analytics, security, and engineering management. Prior to Pivotal, Robert was a co-founder, contributor and VP engineering for Translattice Inc., working on a distributed fully peer to peer PostgreSQL based database product. Previously, Robert was VP engineering for Mu Dynamics (formerly Mu Security) and was senior director of engineering at Symantec after the acquisition of Recourse Technologies. Robert’s career started with 10 years at Motorola Labs working on electromagnetic systems modeling using massively parallel supercomputers, wireless data systems development, mobile security software and e-commerce solutions. He holds several patents in the areas of mobile data, wireless security and e-commerce. Mr Geiger has a Masters of Electrical Engineering degree from the University of Illinois, Urbana and a Bachelor of Science degree in Electrical Engineering from State University of New York at New Paltz.

Talk 2: Apex & Geode: In-memory Streaming on Hadoop

In this session, we will talk about two of the most promising incubating open source Projects, Apache Apex & Apache Geode and how together they attempt to solve shortcomings of existing big data analytics platforms.

Project Apex is an enterprise grade native YARN big data-in-motion platform that unifies stream processing as well as batch processing. Apex processes big data in-motion in a highly scalable, highly performant, fault tolerant, stateful, secure, distributed, and an easily operable way.

Apache Geode provides a database-like consistency model, reliable transaction processing and a shared-nothing architecture to maintain very low latency performance with high concurrency processing.

We will also look at some use cases where how these two projects can be used together to form distributed, fault tolerant, reliable In memory data processing layer.

BIOs

Sandeep Deshmukh

Sandeep is a Senior Engineer at DataTorrent Software India Ltd., working on Apache Apex. After completing his PhD @ IIT Bombay, he has worked in various domains like Petroleum and Life Sciences. Prior to joining DataTorrent, he was Data Scientist at Reliance Industries Pvt Ltd. and Sr. Domain Expert in Persistent Systems Ltd. At Persistent, he was instrumental in porting a DNA sequencing suite of products on Hadoop.

Ashish Tadose

Ashish is a technical lead at Ampool, and worked at PubMatic, as a Lead Engineer, Big Data & Analytics, where he led a team driving large scale data ingestion and real-time streaming analytics solutions. Ashish is experienced in design & implementation of scalable streaming analytics technologies such as Apache Storm, Kafka, Kinesis, Flink, Spark Streaming & Apex. Ashish also delivered data infrastructure to facilitate large scale data ingestion from 6 geographic regions in both AWS cloud and in-prem. Prior to PubMatic, Ashish worked at Verisign as Senior Software Engineer in Big Data Team where he worked on projects which required large scale data processing using Hadoop and MapReduce. Ashish holds Bachelors and Masters degree in Computer Science and passionate about development of products leveraging distributed computing platforms.

Talk 3: Disorderly programming: Avoiding cost of enforcing time order

This talk will cover effects of enforcing time order on performance. Logic which is time order dependent will need some sort of locking to enforce time order. In case of in memory computing and that too distributed in nature, it becomes a problem. The ratio of "cost of an operation on a single node/single shard" to "cost of locking and coherency" is going to be high as compared to traditional disk based systems. This can be analyzed using Little's law and Amdahl's law. Further I would cover techniques to structure program logic to get around ordering and associated cost.

BIOs:

Shripad Agashe

Shripad has more than 17 years of IT experience in executing projects for a broad range of business problems for various large organizations including several Fortune 500 companies. He specializes in Performance and Scalability of compute intensive applications. He is also interested in distributed computing and loves reading "The Morning Paper". He currently works as a tech lead at ThoughtWorks.