Joining a Cluster Server
Clustering Consulting Services
Clustering is the process of joining
multiple servers together to achieve greater fault tolerance or
performance. For years, customers have had the ability to cluster
servers together to add higher fault tolerance for their systems and
applications. However, these clustering solutions were based on
proprietary technology and lacked industry standards, making
applications too costly to develop, manage, and migrate to new
platforms. Because of this expensive proprietary architecture, many
customers were only able to deploy clustering solutions for the most
business-critical applications.
For the latest information about
Joining a Cluster Server
Large Scale Clusters/Grids
Large-scale clusters and
Grids need resilience, operational simplicity,
and management flexibility in their interconnect networks. Building an
interconnect network with separate access and aggregation layers
segments the network, enabling independent maintenance of each segment.
Moreover, this topology provides redundant, load sharing paths between
any source/destination pair. This enables resilient, non-blocking
connectivity between each pair.
Grids are built by
joining multiple
clusters. As a result, they impose additional requirements on
interconnect networks. To adequately provide for user access, Grids need QoS, various SLAs and network security. For IP routing, connections to
the Internet may use BGP; connections between clusters can use IBGP; and
connections within a cluster can use OSPF.
For the latest information about
Large Scale Clusters/Grids
Who uses clustering?
Many different types of organizations use
clustering as a vital
part of the work. A sampling of these include:
-
Marketing:
finding groups of customers with similar behavior given a large
database of customer data containing their properties and past
buying records;
-
Biology:
classification of plants and animals given their features;
-
Libraries:
book ordering;
-
Insurance:
identifying groups of motor insurance policy holders with a high
average claim cost; identifying frauds;
-
City-planning:
identifying groups of houses according to their house type,
value and geographical location;
-
Earthquake
studies: clustering observed earthquake epicenters to
identify dangerous zones;
-
WWW:
document classification; clustering weblog data to discover
groups of similar access patterns.
For the latest information about
Who Uses Clustering
