To all: news from sun
Sun introduces Sun Enterprise Cluster
Solution
PALO ALTO, Calif. -- October 7, 1997 -- Sun Microsystems, Inc.
today announced that it is extending the capabilities of its
award-winning Sun TM Enterprise TM server family with the
introduction of its Sun Enterprise Cluster solution -- which delivers
the next phase of its previously announced Full Moon clustering
Road Map project. Sun's new clustering solution extends the
unique RAS (reliability, availability,serviceability) capabilities of its
servers to provide near-100 percent system availability levels. With
its Sun Enterprise Cluster solution, Sun provides industry-leading
scalability in a four-node cluster, with the ability to apply up to 256
CPUs, 256 gigabytes of memory and 12.8 gigabytes/second of
sustained I/O bandwidth to a single application or a large number
of smaller applications. Sun's new clustering capability enables
customers to dramatically cut network server downtime and its
associated costs while simultaneously allowing customers to
boost performance for business-critical applications.
"Sun's vision is to deliver the "Webtone" for network computing;
that is, provide continuous access to information, from anywhere
to anyone at any time," said John Shoemaker, vice president and
general manager of Sun's Enterprise Servers and Storage Group.
"With Sun Enterprise Cluster technology, we are closer to realizing
our vision than ever before. As we continue to extend the
availability of critical systems and applications on the network, we
expect clusters to move from a niche solution to an integral part of
the enterprise."
The components of the new Sun Enterprise Cluster solution
include:
A high-speed, low-latency cluster interconnect that supports
up to four nodes;
Sun Cluster 2.0 software (which is the next phase of Sun's
Full Moon Road Map project) merges failover and parallel
database functions to maximize availability;
Support for the widest range of cluster-ready SMP servers in
the industry-- adding today the Enterprise 1 and Enterprise
450 workgroup servers and the Enterprise 10000 server;
Support for Sun's full line of storage products, including the
Sun TM RSM Array TM 2000 as well as new products based
on Fibre Channel Arbitrated Loop technology to be
introduced later this year;
The ability to do inter- and intra-domain failover with Sun's
Starfire system;
New service offerings designed to help customers to
maximize the availability and performance levels of their
system.
The Primary Purpose for Clustering: Availability
Clusters are characterized by multiple systems, or nodes, that
work together as a single entity to cooperatively provide
applications, system resources, and data to users. The primary
reason to cluster is for availability: should one node fail, other
cluster nodes can continue to provide data services, and the failed
node's workload can be spread across the remaining servers in
the cluster.
A cluster is only as reliable as the servers, the storage, and the
operating system that make up the solution. According to Roy
Peskoff, research analyst at D.H. Brown, an independent market
research firm, "Sun offers outstanding stand-alone server RAS
capabilities, such as dynamic reconfiguration, among open
systems vendors. Those capabilities, combined with cluster
features such as cascading failover and individual application
failover, places Sun's cluster offering in a strong competitive
position. Additionally, the company's high-end Starfire system, with
its unique ability to cluster system domains, clearly provides open
systems RAS leadership in the data center."
Dynamic reconfiguration improves availability of the individual
nodes within the cluster. The feature originally shipped on Sun's
Enterprise 10000 (a.k.a. StarfireTM) server and will be available on
the Enterprise 3000 to 6000 servers in the Spring of 1998.
Dynamic reconfiguration enables the servers to attach and detach
CPU's, memory and I/O devices with no disruption to the
production environment, tremendously minimizing planned and
unplanned downtime of the single server.
Dynamic System Domains is similar in concept to logical
partitioning in the MVS world. Combining clustering software with
Dynamic System Domains enables a variety of inter- and
intra-domain failover techniques in the Starfire server. This
provides flexibility for data center managers in charge of running
multiple applications -- each with different availability requirements
-- on the Starfire system.
Superior Scalability and Performance
"Unlike NT and other UNIXr competitors, our customers don't need
to cluster in order to meet most required high-performance levels,"
said Andy Ingram, director of marketing for Sun's Enterprise
Server Products Group. "Our strategy is to start with "fat" SMP
nodes which are servers with many processors. This enables our
customers to scale as far as they can in a single server and only
then turn to clustering for additional scalability. For those
customers who want more power than a 64-way Sun server
provides, such as those with extremely large datawarehouses, our
new Sun Enterprise Cluster solution is a superior choice to provide
the additional performance needed."
Today, Sun demonstrated the superior scalability and performance
of its new cluster solution when it announced industry-leading
TPC-C benchmark results (51,871.62 tpmC). The Enterprise 6000
server, in just a two-node cluster configuration, coupled with the
SolarisTM 2.6 operating environment and Sun SPARCstorage TM
Arrays, outperformed the next closest competitor by 31 percent.
(See separate release for further details.)
Delivering on the Full Moon Road Map Project: Sun Cluster
2.0 Software
Sun Cluster 2.0 software is a key component of the Sun Enterprise
Cluster solution. In addition to combining failover and parallel
database functions into the same cluster, new features which add
to the availability and manageability of Sun Enterprise clusters
include: individual application failover, dynamic addition of nodes,
local network adapter failover, and more. (See separate release on
Sun Cluster 2.0 software.)
"The ultimate goal of our Full Moon project is to bring new levels of
manageability to clustering," said Brian Croll, director of marketing
at SunSoft TM. "We will integrate clustering into the Solaris
operating environment, and system administrators, users and
applications will have a single view across the multiple nodes in a
cluster. The net result is that manageability will become a
compelling reason to cluster."
Service and Support
Sun also announced new Sun Enterprise Cluster and Enterprise
10000 Start-Up Services that make it easy for customers to deploy
mission-critical systems. These services are geared to help MIS
managers safely and effectively install and maintain complex
mission-critical systems. The Start-Up services include
installation, consulting and training. In addition, Sun unveiled a new
Campus Clustering service. (See separate release for further
details.)
Pricing and Availability
The Sun Enterprise Cluster solution will be available in October
1997 directly through Sun as well as through the company's
worldwide network of value-added resellers and systems
integrators. The entry price for Sun Cluster 2.0 software, which
includes the Sun NFS TM agent, is $2,000 U.S. Additional pricing
information can be obtained through local Sun sales offices.
Worldwide availability of services will vary by country.
About Sun Microsystems, Inc.
Since its inception in 1982, a singular vision, "The Network Is The
Computer," has propelled Sun Microsystems, Inc., (NASDAQ
"SUNW"), to its position as a leading provider of hardware,
software, and services for establishing enterprise-wide intranets
and expanding the power of the Internet. With more than $8.5
billion in annual revenues, Sun can be found in more than 150
countries and on the World Wide Web at sun.com. |
