hpcs-17-subord

commit 5f06a229c6a651b5b0d6e8c1bd5d6c7bc6227e1e
parent 79068920b16525da87c0f55fcd3349e95311b03e
Author: Ivan Gankevich <igankevich@ya.ru>
Date:   Fri, 24 Mar 2017 20:44:05 +0300

Revise abstract.

Diffstat:
src/abstract.tex
 | 
20
++++++++++++--------

1 file changed, 12 insertions(+), 8 deletions(-)
diff --git a/src/abstract.tex b/src/abstract.tex
@@ -1,10 +1,14 @@
 \begin{abstract}
-	In this paper we describe a new framework for creating a reliable to hardware
-  errors distributed programs. Our main goal was to create a simply yet powerful
-  tool to archiving fault tolerance without creation of checkpoints, memory
-  dumps and other highly disk usage activities. To archive this we first
-  introduce a strong hierarchy of program components (or parts) and then discuss
-  about scenarios for continue computations. The programs parts hierarchy based
-  on Actor model by C. Hewitt, failure scenarios cover most common hardware
-  errors; software error handling are not covered by this article.
+
+In this paper we describe a new framework for creating distributed programmes
+which are resilient to cluster node failures. Our main goal is to create a
+simple and reliable model, that ensures continuous execution of parallel
+programmes without creation of checkpoints, memory dumps and other I/O
+intensive activities. To achieve this we introduce multi-layered system
+architecture, each layer of which consists of unified entities organised into
+hierarchies, and then show how this system handles different node failure
+scenarios. We benchmark our system on the example of real-world HPC application
+on both physical and virtual clusters. The results of the experiments show that
+our approach has low overhead and scales to a large number of cluster nodes.
+
 \end{abstract}