Database Reference
In-Depth Information
Cloud Storage Studio 2: This is a paid tool giving you complete control of your Windows
Azure blobs, tables, and queues. You can get a 30-day trial version of the tool from here:
http://www.cerebrata.com/products/cloud-storage-studio/introduction
CloudXplorer: This is also a paid tool available for Azure storage management. Although
the release versions of this tool need to be purchased, there is still an older version available
as freeware. That older version can be downloaded from the following URL:
http://clumsyleaf.com/products/cloudxplorer
Windows Azure Explorer: This is another Azure storage management utility which offers
both a freeware and a paid version. A 30-day trial of the paid version is available. It is a good
idea to evaluate either the freeware version or the 30-day trial before making a purchase
decision. You can grab this tool from the following page:
http://www.cloudberrylab.com/free-microsoft-azure-explorer.aspx
Apart from these utilities, there are a few programmatic interfaces that enable you to develop your own
application to manage your storage blobs. Those utilites are:
AzCopy
Windows Azure PowerShell
Windows Azure Storage Client Library for .NET
Hadoop command line
To get a complete understanding on how you can implement these programmatic interfaces and build your own
data upload solution, check the link below:
http://www.windowsazure.com/en-us/manage/services/hdinsight/howto-upload-data-to-hdinsight/
Windows Azure Flat Network Storage
Traditional Hadoop leverages the locality of data per node through HDFS to reduce data traffic and network
bandwidth. On the other hand, HDInsight promotes the use of WASB as the source of data, thus providing a unified
and more manageable platform for both storage and computation, which makes sense. But an obvious question that
comes up regarding this architecture is this: Will this setup have a bigger network bandwidth cost? The apparent
answer seems to be “Yes,” because the data in WASB is no longer local to the compute nodes. However, the reality is a
little different.
Overall, when using WASB instead of HDFS you should not encounter performance penalties. HDInsight
ensures that the Hadoop cluster and storage account are co-located in the same flat data center network segment.
This is the next-generation data-center networking architecture also referred to as the “Quantum 10” (Q10) network
architecture. Q10 architecture flattens the data-center networking topology and provides full bisection bandwidth
between compute and storage. Q10 provides a fully nonblocking, 10-Gbps-based, fully meshed network, providing
an aggregate backplane in excess of 50 Tbps of bandwidth for each Windows Azure data center. Another major
improvement in reliability and throughput is moving from a hardware load balancer to a software load balancer. This
entire architecture is based on a research paper by Microsoft, and the details can be found here:
http://research.microsoft.com/pubs/80693/vl2-sigcomm09-final.pdf
 
Search WWH ::




Custom Search