|
Market Roundup November 10, 2006 VMware Launches Virtual Appliance Market Next Generation Clusters from IBM |
|
VMware Launches Virtual Appliance Market
This week VMware announced the creation of a marketplace and
a certification program for virtual appliances. The Virtual Appliance
Marketplace already boasts 300+ virtual appliances available for customers to
evaluate or purchase. Virtual Appliances already available include offerings
supporting collaboration, email security, firewalls, intrusion detection/prevention,
and traffic management as well as basic operating systems. In addition to the
marketplace itself ISVs looking to create virtual appliances can now choose to
certify their appliances through the VMware Virtual Appliance Certification
Program to ensure that their appliances and the ISV itself can fully support
their solutions in this new environment. Certified appliances are currently
available from a range of vendors including Zeus, Zimbra,
Red Hat, ProofPoint, CohesiveFT,
LoadBalancer and PortWise.
It is clear that the use of Virtual Machine software in the
enterprise is moving beyond test and development and into mainstream business
use supporting key front-line business applications. This is verified not
simply by a number of recent surveys but also by speaking with many of the
7,000 or so delegates attending the VMworld 2006. Organizations
now have sufficient trust and experience of virtual machines, especially those
supplied by VMware, to start using them in everyday operations.
Virtual Appliances are designed to help address much of the
complexity still facing organizations as they seek to deploy applications,
whether directly on server platforms or on virtual machines; particularly the
complexity of installation and optimizing operating systems, middleware, and
application software to get a functioning system. In essence a virtual
appliance is a pre-built, pre-configured, and ready-to-run software application
packaged with the operating system inside a virtual machine. All that is needed
is to copy the file containing the virtual appliance to a VMware platform and
it is ready to use. More importantly it has been built and “tweaked” to ensure
that performance is good and that the appliance is reliable. All of the
potential for software stack clashes and conflict are removed as the virtual
appliance is entirely self-contained. In many respects Virtual Appliances are not
simply a software distribution solution but a complete distribution,
installation, and management technology that allows very fast adoption of
mainstream software into production.
If one considers virtual machine software on the x86 platform as still a young technology that is now ready for production use, VMware Virtual Appliance technology is even younger, but it is built on the VMware platform that has earned a fair degree of trust. Overall, we are supporters of the software-appliance approach to computing and we hope that the VMware Virtual Appliance Marketplace will become an important resource for many organizations. However, it is clear that VMware will need to carry out a lot of education and evangelism to ensure that IT organizations understand the Virtual Appliance approach to computing. If organizations learn to understand where they could exploit this architecture, the VMware Virtual Appliance Marketplace could, in time, become a very heavily visited and exploited resource. Indeed, when considered alongside the company’s well known offerings and its other recent developments such as the VMware Virtual Desktop Infrastructure and ACE, Virtual Appliances gives VMware a well rounded set of solutions to sell. It will be fascinating to see how quickly business application ISVs port their offerings to Virtual Appliances. Equally, as such appliances can be deployed and removed easily, we shall watch eagerly to see if customers start asking the ISVs to offer new licensing schemes alongside a Virtual Appliance.
Next Generation Clusters from IBM
IBM has announced its latest generation processor systems
and blades, high-speed connectivity, and coprocessor acceleration technology
for the IBM System Cluster 1350. The new technology has the capability to scale
to 1,024 nodes and is targeted at organizations looking to supercomputing
solutions in a variety of industries including financial services, industrial,
petroleum, and life sciences. The new System Cluster 1350 delivers high levels
of speed, choice, and flexibility to create hybrid supercomputers in a fully
integrated, factory-built, and tested solution. Key server products available
in the new IBM System Cluster 1350 portfolio include: Opteron-based System
x3455, System x3655, System x3755, BladeCenter LS21, and BladeCenter LS41;
Xeon-based System x3550, System x3650 and BladeCenter HS21; Power-based System
p5 505, System p5 505Q, System p5 510, System p5 510Q, System p5 550 and System
p5 550Q and BladeCenter JS21; and as Cell Broadband Engine-based systems,
including IBM BladeCenter QS20. The System Cluster 1350 features networking
options, including Ethernet, Infiniband, and Myrinet interconnects.
HyperTransport (HTx) can be leveraged with
Opteron-based systems thus allowing the 1350 to support the Qlogic/PathScale HTx adapters for memory-intensive High Performance
Computing and Business Performance Computing applications that commonly are
constrained by processor-memory and I/O performance. High-performance
interconnects including 10GBps Ethernet and Infiniband networking from Cisco
and the native 10GBps BladeCenter H Infiniband switch blade, along with Myricom
Myri-10G (10GBps) adapters and switches, as well as double bandwidth (20 GBps) Voltaire Infiniband products such as the Voltaire
Infiniband Pass through Module for Blades, which can eliminate the need for
intermediate switches, are among the interconnectivity solutions supported. The
System Cluster 1350 will also offer ClearSpeed Advance accelerator boards,
PCI-X adaptors designed to improve the performance of numerically intensive
workloads by routing math library routines to the ClearSpeed accelerator board.
IBM System Cluster 1350 pricing varies with the specific configuration.
As mind boggling as it can be, the capability of
state-of-the-art computational prowess packaged up into a standalone server is
sometimes just not enough for some workloads as the demand in many
organizations can easily outstrip this capacity. Although for smaller
deployments a collection of discrete servers can generally meet the workload
demand, for those with serious needs the cluster has always provided a creative
way to build a solution robust yet flexible enough to handle varied workloads
within, and sometimes outside of, organizations. One historic challenge with
clusters was interconnecting all of the components without running into
bandwidth and latency limitations between the interconnects.
We have witnessed substantial improvements in short-distance interconnects in
recent times, and technology such as the Voltaire Infiniband Pass-through
Module. The ability to offer 20 GBps throughput over
longer distances is a far cry from the 10MBps Ethernet or even 100MBps FDDI we
had a scant decade ago. Alternatively, HTx provides a
chip level interconnectivity with throughput that would have been found only in
the dreams of science fiction a generation ago. Of course, on top of all of
this is the sheer computational horsepower of the System x, System p, or
BladeCenter blades: almost an embarrassment of riches to offer to the cluster-oriented
customer.
Technically there is a lot here to appreciate, but technology does not sell itself; it must support a solution that the market wants to buy. Obviously, the sheer scaling potential of the 1350 is impressive, but there are a relative few who need such a solution. However, scale works both ways, and a cluster is not necessarily made of 1024 or even 512 nodes; it is quite viable with five or ten. The challenge for IBM is to deliver a family of clustered offerings that target specific needs, perhaps with a vertical orientation, that clearly articulate an answer to problems that organizations are facing and will be easy to understand. In some respects, the answer to customer problems needs to be simple and direct, even if it is implemented through a relatively complex clustered solution. While Big Blue, and the industry overall, have gotten better at engaging in business-focused as opposed to technology-focused discussions, clusters have a high geek factor, yet as with most things geek they ultimately experience their highest success and market potential when the technology takes the supporting role to the business discussion, as opposed to the other way around.
Microsoft and Novell: Bambi Meets Godzilla?
Microsoft and Novell have recently announced a set of
agreements to build, market, and support a series of new solutions to make
their products work better together. Microsoft will now recommend SUSE Linux
Enterprise for customers who want Windows and Linux together. The two companies
are going to create a joint research facility at which Microsoft and Novell
technical experts will architect and test new solutions with customers and the
open source community, including virtualization solutions, Web services and
service-oriented architectures (SOA), and interoperability between office
productivity applications. They have also promised to provide each other’s
customers with patent coverage for their respective products. Microsoft and
Novell will each provide covenants that promise not to assert their respective
patent rights against customers who have purchased or licensed products from
the other. Microsoft has also made some agreements with Novell regarding the
open source community. Finally, the two companies will pursue a variety of
joint marketing activities to promote the adoption of technologies on which
they’ve collaborated. Novell has posted a FAQ on its web page for the open
source community explaining details of the agreement as it relates to the
community.
The implications of what Microsoft and Novell are up to
should keep pundits busy for awhile. Certainly a flotilla of
lawyers are earning their holiday bonuses this year. But we’re more
interested in the why of it all, and more importantly, why now? Novell has been
losing ground to Red Hat and others despite the fact that many believe SUSE to
be the superior distribution from a technical viewpoint and despite the fact
that companies like IBM have made significant financial contributions to make
sure Novell could do the right things with SUSE. At the same time, Microsoft is
turning up the heat as Vista hit RTM (release to manufacturing) this week,
surrounded by discussions of changes to the wording of its upgrade policy,
concerns over what features are available on which versions of Vista, and what
exactly the WGA (Windows Genuine Advantage) will be like under Vista and how
likely this is to turn customers off to Microsoft and cause them to seek
alternatives. Certainly both companies are vulnerable. Novell has had many
recent changes in executives and strategy and has yet to find a compelling
message for its mix of open source SUSE and traditional Novell products. Microsoft’s
announcements, changes, and re-announcements around Vista have led to the
development of a web community akin to that which follows the television series
Lost, as the curious try to figure
out exactly what is really going on and what it all actually means.
On another front, patents come up frequently in these agreements. Said agreements are not irrevocable, and they are finite, but they certainly indicate an intent and a direction from the two companies. It is a sad statement on the state of the industry that more and more companies are making money from patents and intellectual property not by making products with them but by threatening other vendors who may be infringing those patents. SCO has certainly become the poster child for how not to do this, but respectable Linux supporters like HP came out early indemnifying its customers who used Linux in case the threat might be credible. As far as we know, no Linux customers have been sued for using a distribution, but vendor versus vendor could be the subject of the next Celebrity Deathmatch series on MTV. In this particular instance, although Microsoft and Novell have promised to play nice, many disgruntled folks in the open source community aren’t sure if what Novell has done is actually in violation of the GPL or whether they’re just selling out. This announcement has given the press a bit of excitement while they wait for the Zune and Vista launches. But in terms of real impact on users? Well, we’re not holding our breath.
HP and World Wildlife Fund-US (WWF-US), known for their use
of a Panda as their logo, recently announced a joint worldwide initiative to
reduce HP’s greenhouse gas emissions from its operating facilities, to focus on
educating and inspiring others to adopt best practices, and to use HP
technology in conservation efforts around the world. Among the jointly stated
goals is that by 2010, the carbon dioxide emissions from HP-owned and HP-leased
facilities will be reduced by up to 15% below their 2006 levels. This will be
achieved by both WWF-US and HP working together to identify the best technology
and practices available to reduce energy use in these facilities. Another
stated goal is that HP will ramp up its efforts to explore and purchase
cost-effective renewable energy. Yet another goal is for HP to develop energy
efficiency measurements for its products. WWF-US and HP will also work together
to define and implement strategies to address and educate individuals and
organizations about climate change, as well as to bring to light the top best
practices adopted by both consumers and businesses. The two companies have
stated they will also leverage HP technology to advance the science and ideas
of adapting to the coming climate change. One of their initial joint projects
also involves using funding from HP to study the effects of climate change on
the ecosystem of North America’s Bering Sea.
In our industry’s ever-steady march toward eco-friendly
policies and products, we believe that HP has taken another step in the right
direction. In the famed report released by Greenpeace earlier this year, HP
ranked sixth out of fourteen companies surveyed, somewhat behind Dell and
Nokia. Not a bad showing, although none of the companies surveyed passed as
being totally green. Dell and Nokia both earned a score of 7 out of 10, and as
anyone who has gone to school knows, a 70% may be
passing but it’s hardly spectacular. However, by implementing these latest
policies and procedures, as well as by working with the World Wildlife Fund-US,
next year’s Greenpeace ranking (if Greenpeace does do another survey next year)
should up HP’s score from its current 4.7 and move the company closer to the
green. A current article on Greenpeace’s website acknowledges efforts by HP to
“green up” their practices.
This is great news for HP. When a formerly adversarial
company publishes an article endorsing HP as one of the leaders in the industry
in taking action against toxic chemicals, the sales to environmentally
conscious consumers (or those that are just saying they are environmentally
conscious for the good press it generates) are sure to follow. However, it is
perhaps not really a shocker that when HP decided to team up with an environmental
group, they choose the WWF-US rather than the more adversarial Greenpeace. Greenpeace’s
website is full of announcements and denouncements of tech companies and their
poisonous policies. WWF-US, on the other hand, only provides one article about
the tech companies: the one announcing the partnership with HP. Seems that
sounding the warning alarm doesn’t gain an organization any popularity points.
But who cares? As long as companies are moving forward in implementing environmentally friendly policies and practices, no matter how slow that forward motion is, it doesn’t matter which conservation organization wins the pińata. We all eventually end up with the prizes of clean(er) water, clean(er) air, and amazing biodiversity.