Akamai Selects Parallocity’s ZVM-U Dynamic Software Analysis Framework

  Akamai Selects Parallocity’s ZVM-U Dynamic Software Analysis Framework

ZVM Software Leveraged for Defect Detection in Regression Test and Production
                                 Environments

Business Wire

SANTA CLARA, Calif. -- March 5, 2013

Parallocity, a key player in the Automated Software Quality (ASQ) industry and
provider of the Zeus Virtual Machine® (ZVM) Dynamic Analysis Framework for
automated defect detection and code analytics, today announced that Akamai
Technologies, Inc., the leading cloud platform for helping enterprises provide
secure, high-performing user experiences on any device, anywhere, has licensed
ZVM-U for software defect detection during development as well as on
production servers used in Akamai’s Intelligent Platform for online business
management.

The ZVM Dynamic Analysis Framework forms the core Parallocity’s ZVM-U and
ZVM-K products. ZVM is unique in the industry because it incorporates a
compiler front-end to analyze an application’s source code before producing an
instrumented image for runtime analysis. This Intelligent Instrumentation™
process enables ZVM to make smart decisions about which data structures to
monitor during application execution. As a result, ZVM-instrumented
applications run orders of magnitude faster in a fraction of the memory
footprint than applications instrumented with traditional dynamic analysis
tools. ZVM typically requires just 10% to 20% more system memory during
runtime and ZVM’s impact on application performance is just 1.2X to 5X
compared to slowdowns of 20X to 300X for traditional dynamic analysis tools.

“Parallocity’s ZVM-U error detection framework will be a useful tool in
helping Akamai continue to deliver high-quality end-user experiences for our
customers,” stated Eric Crawley, Senior Director, Web Experience Engineering
at Akamai Technologies. “Because ZVM-instrumented images are so fast and
memory efficient, we can run them selectively on our production servers to
find errors that would have otherwise been very difficult to detect.”

Akamai’s Intelligent Platform manages the underlying complexities of running
an online business, from device and format proliferation, to application and
network security, to performance and reliability issues. The Platform is made
up of over 127,000 servers deployed in 81 countries, which are all controlled
by Akamai software that constantly monitors Internet conditions.“From the
start we designed ZVM to work in the regression test environment. That meant
incorporating algorithms and heuristics to enable ZVM to scale without
creating a major drag on performance or significant memory bloat,” stated
Himanshu Shukla, CTO of Parallocity. “Today we have customers running ZVM
seamlessly on a range of platforms from small, embedded devices with just 64MB
of system memory all the way up to powerful many-core platforms with 72GB of
memory running complex applications with millions of lines of code and dozens
of concurrent processes.”

ZVM is also unique in the industry because it incorporates algorithms and
heuristics that are thread schedule agnostic and timing independent. As a
result, ZVM can find data races, deadlocks, and stalls in complex
multi-threaded code with high precision. Further, because it is a runtime
tool, ZVM reports memory errors and memory leaks with 100% accuracy and zero
false positives. As such, developers can easily identify the root causes of
reported defects and fix them quickly.

“We are pleased with Akamai’s selection of ZVM-U to improve the quality of
their cloud-based platform and their end-user experience,” stated Shekhar
Ambe, CEO of Parallocity. “Akamai’s use of our technology in its production
environment shows the power of ZVM-U in finding hard-to-detect errors and
under circumstances that no other tool in the industry can replicate.”

About Parallocity

Parallocity is a key player in the $2B Automated Software Quality (ASQ)
industry. Its Zeus Virtual Machine (ZVM) Dynamic Analysis Framework automates
the detection of hard-to-find concurrency defects, memory errors, and memory
leaks while providing comprehensive analytics for highly complex, single- and
multi-threaded applications, loadable kernel modules, and device drivers. For
more information please visit www.parallocity.com.

Parallocity is a trademark and Zeus Virtual Machine is a registered trademark
of Parallocity, Inc. and may be used only with permission and proper
acknowledgement. All other brands, product names, and trademarks are property
of their respective owners.

Contact:

Parallocity, Inc.
David Cowan, 408-396-3100
david@parallocity.com
 
Press spacebar to pause and continue. Press esc to stop.