Steve Vinoski, Verivue

Steve Vinoski is a member of technical staff at Verivue, a startup in Westford, MA, USA. He was previously chief architect and Fellow at IONA Technologies (now part of Progress Software) for a decade, and prior to that held various software and hardware engineering positions at Hewlett-Packard, Apollo Computer, and Texas Instruments.

Over the past two decades Steve has authored or co-authored approximately 80 highly-regarded publications on distributed computing and enterprise integration for magazines such as IEEE Internet Computing, C/C++ Users Journal, and C++ Report, and co-authored the book "Advanced CORBA Programming with C++" with Michi Henning, published in 1999 and regarded by many as "the CORBA bible." He started working with distributed computing systems at Apollo Computer in 1988.

Presentation: "Let It Crash...Except When You Shouldn't"

Track: When Things Break

Time: Thursday 14:05 - 15:05

Location: St James’s Suite, Fourth Floor

Abstract:

Erlang is well known for its "let it crash" philosophy for dealing with errors, where handling specific errors is generally avoided in favor of having hierarchical process supervisors restart any failed processes. This works surprisingly well, allowing developers to pay little attention to error handling and yet still create Erlang applications providing high availability and uptime.

Still, blindly just letting things crash can sometimes result in undesirable client-visible service failures and outages. Achieving even higher levels of availability means avoiding crashes for some types of errors and deciding specifically when and how to crash for the rest. In this talk, Steve takes a detailed look at common types of Erlang errors that can be avoided and how to steer clear of them while still adhering to the "let it crash" philosophy.