Special Features:
Avaki SUPPORTS COMMERCIAL GRID
TERRITORY WITH JAVA MAKEOVER
By William Fellows, the451.com
To create a greater opportunity for its software in the commercial sector,
grid specialist Avaki has made over its Data Grid product with Java 2,
Enterprise Edition (J2EE) in release 3.0. Up until now, it had used C++. Avaki
is hoping to make its grid-based data access software more appealing outside
of the high-performance computing world.
The move comes at a time when grid companies are beginning to turn their
focus
toward commercial engagement opportunities, believing the technology is now
maturing and gaining enough visibility.
The message Avaki hopes a Java version of its Data Grid software will be
attractive to enterprise IT groups in commercial organizations, and believes
it's now time to focus on this opportunity. It has also created better
management tools and integration with widely used enterprise directory
systems. A CIFS implementation will reach out to Microsoft shops.
Competitive landscape While Avaki's compute software can connect users with
resources on a grid, its core functionality is providing access to data.
Companies such as Platform and Entropia that offer compute grids are therefore
more complementary than competitive in theory, but Avaki is expanding its
focus with this new release into where these companies play.
IBM has created perhaps the best-known instance of a data grid using its
own
DataJoiner software. Sun also says it's created data grids using its Grid
Engine software. Ironically, IBM, Sun and HP are all Avaki partners.
Avaki is the latest company to begin targeting the commercial opportunity
for
grid technology. Most haven't done much except throw some marketing behind
existing products and services. Avaki's makeover in Java is a significant
development and will elevate its whole proposition and that of grid computing
in general.
Technology Avaki promised last summer that it would be providing greater
integration with Java. The 3.0 release of its software for accessing data
stores over grids has been recast in Java, although the compute grid software
which runs on top remains unchanged. Using Java makes Data Grid more
accessible to enterprise application developers and the entire J2EE
community.
Version 3.0 also includes more powerful file-based access to data, higher
availability and failover. Avaki has extended the authentication mechanism so
that LDAP, Active Directory and other directory mechanisms can be used to
integrate and manage data grids created using its software. There's also a new
GUI-based system manager.
Typically, administrators prefer to retain older command-line management
tools
because they're easier to use, maintain and extend, but Avaki says all of the
users it's tested the new manager with have preferred it to a command-line
interface.
Data grid is a packaged software application that enables customers to
access
files and data regardless of location, administrative domain or platform.
Avaki likes to claim it can deliver wide-area data access with local-area
performance. The important piece, however, is that it operates beyond
firewalls and across multiple sites. It creates a federated data service that
securely serves data from existing sources to users and applications over a
wide-area, leaving the data in place.
Data grids connect organizationally and geographically diverse users,
files,
and databases through a coherent network-wide data infrastructure. IBM's
recent e-Diamond project linking five UK hospitals and enabling them to share
mammogram images and data is perhaps the best-known example of a data grid. It
uses Mirada software and IBM's own DataJoiner software. Sun claims to have
implemented a similar data grid at hospitals in the US.
Avaki Data Grid 3.0 is available as a stand-alone offering, or bundled with
Avaki's Compute Grid 2.6. Although Avaki won't be making over Data Grid in
.NET, it says .NET can be used as readily as XML, Java or other Web services
component languages to access resources over its software. Sun's Grid Engine
can be used against Data Grid, for example. Next for Avaki is implementing
CIFS access to resources via its software. Currently, access is only supported
from NFS calls. CIFS support is due by midyear 2003.
Context While Avaki has largely endorsed IBM's world view of grid computing
as
virtual, distributed and complex -- versus Sun's more simplistic 'clustered'
interpretation by default -- the debate between different camps over what does
or doesn't constitute a grid is ongoing. Indeed, Avaki founder, CTO and grid
pioneer Andrew Grimshaw has seen fit to lend his own comments recently in
GridToday.
While different vendors use different terminology to describe the various
levels of grid implementation and opportunity -- such as cluster, compute,
data
enterprise, partner, service, public and global, to cite just a few --
Grimshaw
says a grid can be a more general term of classification than any of these.
"This is important because to realize their full potential, grids cannot be
limited to any one of these," he says. "A grid is a collection of distributed
resources connected by a network, possibly at different sites and in different
organizations."
"What distinguishes these resources is that they have a network interface
and
some software that grid-enables the device. Thus, one could say that from a
hardware perspective, potential grid resources range from toasters to
teraflops. One could argue that the above definition of grid is what we used
to call a 'distributed system.' I do not dispute that it is what we used to
call a distributed system. To me, grids are the evolution of distributed
systems to a wide area, multi-organizational context."
Grimshaw says that grids, by their nature, have the potential to become
very
complex, making complexity management the number one priority of any grid
middleware system. From a user perspective, he says, grids are all about
resource sharing and virtualization across sites and organizations.
Business model Avaki readily admits that technology such as its Data Grid
software will, in future, be rolled up as part of much broader grid products
and portfolios, although it has no intention of relinquishing control of its
own destiny. Indeed, it's in the middle of raising more funding that will be
used to drive international expansion. It has raised $16m so far.
Avaki says it now has at least eight customers that have deployed its
software
in production environments. Typically, users are testing its software in
projects linking two or three resources and spending "tens" of thousands of
dollars. If they decide to implement the software in production environments,
the level of spending jumps to over $100,000, Avaki says.
Although something of a pioneer in applying Java to grid technology, Avaki
won't comment on why, unexpectedly, a new Java committee wasn't created at the
recent Global Grid Forum meeting. Java is well understood, and there are other
things more critical to the development of the grid community. The GGF's focus
is on Web services access to grid services.
Competition While Avaki's compute software can connect users with resources
on
a grid, its core functionality is providing access to data. The company
estimates that most users are implementing its software for data access and
that 50% are also using the compute functions. Companies such as Platform and
Entropia that offer compute grids are therefore more complementary than
competitive.
IBM has created perhaps the best-known instance of a data grid using its
own
DataJoiner software, and with Platform as a partner, this probably represents
the biggest threat to Avaki. Sun also says it's created data grids using its
Grid Engine software. Ironically, IBM, Sun and HP are all Avaki partners.
Avaki has established a leadership position in data grids and is tapping
into
a commercial opportunity with Java. Avaki will need to demonstrate that it
can turn the new technology into reference commercial customers, of which
there are few right now. Opportunities Threats There's a growing consensus
among grid companies that a commercial opportunity is emerging. Partners such
as IBM and Sun could end up being competitors, and commercial organizations
may not be ready to adopt grids yet.
|
