Software Engineering Radio

Recording Venue: Oracle Java One 2010, San Francisco
Guest: Nati Shalom

Host: Robert

In this episode, Robert talks with Nati Shalom about the emergence of large-system architectures consisting of a grid of high-memory nodes. As memory has become faster and cheaper more rapidly than has disk storage, application state has increasingly migrated to memory in the form of caches. Memory grids take this one step further, using memory as the system of record and disk is used as a write-behind journal for recovery purposes.  Data and processing are colocated on the same node, which hosts what would be multiple tiers in the popular multi-tiered architectures. Data partitioning and distributed algorithms such as map-reduce become critical design decisions. Nati also discusses reliability and availability considerations of memory grids.

Links:

• Nati Shalom’s blog
• Nati Shalom’s slides on slideshare
• Gigaspaces
• Nati Shalom on Twitter
• Designing a Scalable Twitter
• RAM is the new disk
• The Case for RAMClouds by Ousterhout et. al.
• Are Cloud Based Memory Architectures the Next Big Thing?

Podcast: Play in new window | Download

Recording Venue: MongoSF 2010
Guest(s): Dwight Merriman

Host(s): Robert
Dwight Merriman talks with Robert about the emerging NoSQL movement, the three types of non-relational data stores, Brewer’s CAP theorem, the weaker consistency guarantees that can be made in a distributed database, document-oriented data stores, the data storage needs of modern web applications, and the open source MongoDB.

Links:

• 10-gen
• NoSQL – the Shift to a Non-Relational World by Dwight Merriman
• A series of blog posts on consistency in a distributed database (see entries from March 26 to April 15, 2010)
• Slides and video from MongoSF 2010
• NoSQL hub
• NoSQL mailing list on Google
• Follow Dwight Merriman on twitter
• 10-gen’s blog
• Brewer’s presentation on the CAP theorem
• MongoDB

Podcast: Play in new window | Download

Recording Venue: QCon
Guest(s): Jay Kreps

Host(s): Robert
Jay Kreps talks about the open source data store Project Voldemort. Voldemort is a distributed key-value store used by LinkedIn and other high-traffic web sites to overcome the inherent scalability limitations of a relational database. The conversation delves into the workings of a Voldemort cluster, the type of consistency guarantees that can be made in a distributed database, and the tradeoff between client and the server.

Links:

• Project Voldemort
• Jay Kreps presentation at QCon San Francisco 2009
• Google mailing list for Project Voldemort
• Project Voldmort on github
• LinkedIn blog entry on Project Voldemort
• Amazon’s paper on Dynamo
• NoSQL hub
• Google mailing list for NoSQL

Podcast: Play in new window | Download

Recording Venue: QCon San Francisco 2009
Guest(s): Philip Zeyliger

Host(s): Robert
Philip Zeyliger of Cloudera discusses the Hadoop project with Robert Blumen. The conversation covers the emergence of large data problems, the Hadoop file system, map-reduce, and a look under the hood at how it all works. The listener will also learn where and how Hadoop is being used to process large data sets.

Links:

• Cloudera
• Cloudera Hadoop distribution
• Philip’s E-mail
• Presentation slides
• Book: Hadoop – The Definitive Guide
• Hadoop Project

Podcast: Play in new window | Download