This is an excerpt from the bestselling book
Oracle Grid & Real Application Clusters. To get immediate
access to the code depot of working RAC scripts, buy it
directly from the publisher and save more than 30%.
There is always a lingering
doubt in everyone?s mind as to what Grid is exactly and how it
differs from other well known architectures such as Cluster and P2P
architectures. In this chapter, the true nature of Grid and
Clustering architectures will be highlighted.
There is a somewhat hazy notion
and understanding of Grid and the types of Grids among many IT
technologists. Sometimes they do not find a clear demarcation
between grids and other related technologies like clusters. Those
differences between the Clusters and Grid will be examined.
Another thing that often needs
explanation is the concept of On-Demand and Utility Computing Model,
which is often termed as Grid Computing. The on-demand and utility
computing buzzwords will be explored as well.
The Clustering of Servers has
been around for about two decades and it is one of the most widely
understood, being deployed in scientific, research and commercial
worlds. DBAs are well versed with the concept of aggregating the
servers, and view them as the single system image (SSI). Now, with
the gradual adoption of grid technologies, one often wonders if grid
works with clusters or if it replaces the clustering. How does grid
change the whole perspective of the sharing of the servers?
The Nature of Grid
Grid Computing is an emerging
infrastructure that aims at providing a mechanism for sharing and
coordinating the use of diverse computing resources. The word Grid
is often used as an analogy with the electric power grid, which
provides access to electricity.
As Grid Computing makes long
strides and impacts many organizations in the IT world with its
utility-like access to computational resources, a question remains
in everyone?s mind if Grid Computing can become similar to the
Electric Power Grid of the 20th century. While the state of Grid
Computing is still in its infancy, there are definite signs of
similarities according to Rajkumar Buyya and Madhu Chetty, the
researchers from University of Melbourne, Australia.
As seen in Table 2.1, there are
many apparent similarities in the electric power grid infrastructure
and computational grid structure. However, the computational grid is
more varied and more complex than an electric grid. The existence of
hardware components and user specific software components and
applications make all the difference. Nevertheless, the comparison
is worth noting.
ELECTRICAL POWER GRID
COMPUTATIONAL POWER GRID
Heterogeneous: thermal, hydro,
wind, solar, nuclear, others
PCs, workstations, clusters, and
others; driven by different operating and management systems
Transmission lines, underground
cables. Various sophisticated schemes for line protection.
Internet is the carrier for
connecting distributed resources, load, and so on.
Bus Energy transmission Voltage
Node Computational transmission
Power station (turbo generators,
hydro generators), windmill
Grid resource (computers, data
sources, Web services, databases)
devices: for example, mechanical energy for fans electricity for
TVs, heat for irons
Heterogeneous applications: for
example, graphics for multimedia applications, problem solving for
scientific or engineering applications
Security / safety
Fuses, circuit breakers, and so
infrastructure, and PKI-based grid security
Only storage for low-power DC
No storage of computational
power is possible.
Advanced metering and accounting
mechanisms are in place
Local resource management
systems support accounting. Resource brokers can meter resource
Many standardization bodies
exist for various components, devices, system operation, and so on.
(For example, the IEEE publishes standards on transformers,
harmonics, and so on.)
Forums such as Global Grid Forum
and the P2P Working Group promote community practices. The IETF and
W3C handle Internet and Web standardization issues.
Source: Rajkumar Buyya and Madhu
Table 2.1 Electrical and
computational Power grids: A comparison.
With sudden interest in the grid
and grid-related technology, many IT vendors and analysts are
creating their own vision, definition, and solutions in the grid
space. Grids have moved from the obscurely academic to the highly
popular. We read about Compute Grids, Data Grids, Science Grids,
Access Grids, Knowledge Grids, Bio Grids, Sensor Grids, Cluster
Grids, Campus Grids, Tera Grids, and Commodity Grids as proposed by
various IT companies and researchers. There are so many flavors and
variations, based on the functionality and sometimes based on the
understanding. Many IT vendors freely term their solutions as Grid
technologies and try to fit them into some category.
The following is a quote from
?Ultimately the Grid must be
evaluated in terms of the applications, business value, and
scientific results that it delivers, not its architecture.
Nevertheless, the questions above must be answered if Grid computing
is to obtain the credibility and focus that it needs to grow and
Here is Ian Foster?s 3-point
checklist for a grid computer:
Coordination of Distributed
Resources - Grid controls and integrates
different resources and users within different control domains ? for
example desktops versus large computers, different units of the same
enterprise, and different enterprises. It also addresses the issues
of security, policy, membership, and payment.
Using Standard Pen, General
Purpose Protocols, and Interfaces - Grid
Computing is based on various protocols and interfaces. These
protocols and interfaces control the authentication, resource
discovery, and resource access.
Quality of Service
- Grid aims at delivering at non-trivial quality services in terms
of response time, throughput, availability and security. This is the
motivation for the community to move towards the grid-computing era
and meet the ever-increasing user core application demands. Grid
computing becomes more of a utility from the user?s perspective.
Ian Foster?s 3-point checklist
provides a broad guideline. With these definitions in mind, many IT
vendors are coming up with varied Grid Architecture solutions. At
the same time, some vendors are pitching their cluster solutions.
This happens because the clustering or sharing servers concept and
clustering solutions loosely fit into the broader sense of grid