Dell (DELL) said the acquisition will allow it to offer a broader range of IT services, and a built-in market for its hardware among Perot Systems’ existing customers. Perot Systems has a strong footprint in the health care and government, which provide 48 percent and 25 percent of its revenues, with the remainder in the enterprise sector. Those two sectors are expected to see strong growth due to the Obama administration’s economic stimulus plan, which is expected to boost the adoption of electronic health records and upgrades of federal agencies’ IT infrastructure.
Miller in a separate Twitter message says this deal is designed to boost enterprise and data center businesses. Other tweets and analysts suggest this is to counter the HP and EDS merger, as well as IBM, to build out higher-margin IT services and outsourcing businesses.
This goes along with what I heard through the grapevine—that Dell has been working hard to create a new data center service business. My source tells me that a person at Dell who was promoted to head a new group is traveling all over the country and cannot be easily contacted. Obviously, he is aligned with this movement.
Rich Miller of DatacenterKnowledge.com wrote an article on the recent server market.
Due to the economic slowdown, the overall server market did not do well. It is quite understandable. (The following discussion is during the forth quarter of 2008 unless otherwise specified. )
However, I found some of the statistics interesting. For example:
The unit shipment is down by 12% and the total revenue went down by 14% to $13.5B in comparison to a year ago
The server market share is as follows: IBM (36.3%), HP (29%), Dell (10.6%) and Sun (9.3%)
In spite of the slowdown in the server market, blade servers are doing well. The revenue went up by 16.1% while the shipment went up by 12.1%
Although the market share varies from time to time, these four vendors dominate the server market. In total, their shares occupy 85.2% of the entire server market. The reason for blade servers’ surge is attributed to energy efficiency.
There are two aspects of blade servers. On one hand, the support argument for blade servers is energy efficiency because of sharing of PSU (power supply unit) and cooling (fans) among several blade servers. I blogged on this in my previous posting. On the other hand, there is an argument against it as I blogged in here.
Now which argument is correct? That depends , as I mentioned in my recent blog. There are several arguments like this in the data center space like DC vs AC power delivery. Depending upon assumptions and contexts, a different conclusion is derived. It is important to know what we are arguing with clear understanding of assumptions and contexts.
One of my long-term interests, which is true of our firm AltaTerra in general, is to try to contribute to these discussions in a way that enables meaningful comparisons, though generally this is not by using formal standards.
The main business of Dell is selling server and other IT gear. Agam Shah of Networkworld wrote this article to explain how Dell plans to conduct its business to help its customer conserve energy with their gear.
If you are a server vendor, there are two major ways to conserve energy:
Produce energy efficient gear (and set an appropriate refresh cycle for product lines), and
Provide virtualization technology and support
Dell now claims that a short refresh cycle of 3 years instead of historical 7-8 years makes sense as Albert Esser, vice president of Dell's data center infrastructure group says:
“"It is very clear that refresh rates of three years ... the total cost of ownership is less than keeping your old machines running," Esser said. Old servers unable to distribute workloads could utilize excess power and lead to system administrators buying exponentially expensive servers to fill IT needs.”
As I indicated before you need more memory to support virtualization:
“"Virtualization requires an awful lot of memory, but doesn't require an awful lot of CPU horsepower," Esser said. Dell plans to load two-socket servers with memory attachments typically found in 4-socket servers, Esser said. Similarly, four-socket servers will have memory attachments typically found in eight-socket servers.”
As for major virtualization technologies, Dell covers them all as:
“The company already ships products that are enabled for VMware, Microsoft's Hyper-V and XenServer hypervisors, a thin piece of software that unites multiple operating systems on a server and allocates resources accordingly.”
Other areas Dell works to save energy include power supply units to be more energy efficient.
It is good to know that a server vendor like Dell is keen on energy efficiency. One interesting thing that came to my mind is that as in my previous blog, an air-economizer system may shorten a server’s life because of unfiltered harmful gasses and dusts. One argument is if the server life is only three years and it gets refreshed every three years, the impacts of those undesirables may be within a reasonable tolerance level. So, from that point of view, the shorter refresh period is good.
As we expand the coverage of this blog from green data centers to green IT, I note an article that illustrates some of the relationships between the two topics—and how large and complicated "green IT" can be. The article appeared at the end of 2008 in the Wall Street Journal.
The article analyzed Dell’s declaration of “carbon neutrality” and concluded that Dell’s claim is not valid.
It defines “carbon neutrality” and “carbon footprint” in passing:
“The term may suggest a company has reengineered itself so that it's no longer adding to the carbon dioxide and other greenhouse gases scientists say are contributing to climate change.”
“The amount of emissions Dell has committed to neutralize is known in the environmental industry as the company's 'carbon footprint.'"
As the authors rightly point out, the definition of “footprint” is not standardized:
“there is no universally accepted standard for what a footprint should include, and so every company calculates its differently. Dell counts the emissions produced by its boilers and company-owned cars, its buildings' electricity use, and its employees' business air travel.”
Their problem with Dell’s claim consists of two areas:
Some items are not included in “carbon footprint”, such as “the oil used by Dell's suppliers to make its computer parts, the diesel and jet fuel used to ship those computers around the world, or the coal-fired electricity used to run them.” This is “about 10 times the footprint Dell has defined for itself.”
The purchase of “carbon credits” by Dell is not really reducing the company's footprint.
The article takes a negative view of Dell’s claim. I, however, consider Dell commendable for taking a stance at this early stage when carbon neutrality is not well defined. Somebody should start from somewhere. It is not completely fair to single out Dell when Yahoo and Google, among others, are taking a similar approach.
Dell has taken a first step towards the goal of “carbon neutral.” Now Dell and others should begin to develop a more precise and standardized definition.
Bridget Botelho of Techtarget reported that in addition to Sun, Microsoft and Hewlett Packard, Dell has entered data center consulting services.
Their services include:
“to help users extend the life of existing facilities and avoid building out new data centers.”
In order to satisfy today’s computing resource demands, there are usually only two solutions according to Dr. Albert Esser, the vice president of power and infrastructure solutions at Dell.
build a new data center or
renovate an existing one.
Both solutions entail very high cost.
“Dell's data center services strategy is threefold: extending the life of existing data center by implementing server virtualization to increase server utilization, decommissioning out-of-use equipment and refreshing legacy systems.”
The first two strategies are self-explanatory, but what is “refreshing legacy systems”? This is a fancy way of saying dumping old servers and getting new (Dell) servers. This is a clever way of pushing Dell’s servers although I agree that newer servers tend to be more powerful and power-efficient than many servers in action.
In addition, Dell advises their clients, to do things such as:
Raising the data center temperature,
Hot/cold aisle containment, and
Use PUE to optimize a data center under service
An interesting thing is that these bits of advice are based on retrofitting of their own data centers with such things as virtualization. Any advice based on real experiences would be very convincing.
Two recent articles in GigaOM further indicate advancement of Cloud Computing or server side computing. One is an article regarding Dell, which
appears to be changing its major business of manufacturing and selling of PCs to something else, as
Wall Street Journal reports.
It speculates that Dell is talking to contract manufacturers to sell
their manufacturing plants. Dell may be moving from the current major business
to something totally new like Cloud Computing. Dell currently has much smaller
form-factor PCs and those may be a device for starting business in Cloud Computing.
Dell could sell hardware for Cloud Computing and services related/required by
Another article discusses the death of the DVD sales business in
Korea, which is much more connected and on-line than most of countries. Its
broadband penetration rate is 90.1%. In Seoul, more than 50% of people download
movies from the net, and renting and sales of DVDs nose-dived.Because of this, Sony Pictures is leaving
the Korean market like other Hollywood studios.Even VoD is not replacing it but by cloud. People there are using Web
based storage called Webhard instead.
In addition to the two news, Microsoft has announced Midori for
cloud computing and recently, Google released Chrome. Chrome is not intended to
browse Web pages but execute applications in the cloud.Those collectively reinforce the market
trend of moving towards Cloud Computing.In
addition to simply replacing functions and applications on the client with the
server side, new services and functions could be generated. Storage in the
cloud was mentioned a few years ago but there was no much market then.But as the second article reports,
such business is developing and expanding.
What this means is that computing is shifting towards the server
side, demanding more computing power and storage capacity at data centers.This will encourage the development of
smaller yet more powerful chips and the packing of more blade servers in a
chassis.In turn, this will require
more (huge) data centers and more power to run them.If enough number of power plants are not developed in time to accommodate those rising
power requirements, demands cannot be satisfied.
Greening data centers have started with low hanging fruits but as Cloud
Computing moves into a center stage, more complex and sophisticated
technologies and policies will be required, including renewable energy and so
on. I will discuss this in a future blog.
Posted By Zen Kishimoto,
Friday, September 5, 2008
Updated: Sunday, September 7, 2008
As discussed in many
publications, server virtualization is easy to understand since it
allows an operator to consolidate multiple servers to a single machine
without losing any services. Although there are a few variations, the
way virtualization is defined and implemented for servers is pretty much
well agreed upon and known. In a legacy server setup, on top of a bare
machine (hardware), there is a host operating system (OS), such as
Windows and Linux. Then, a single or multiple applications execute
utilizing the host OS. In case of virtualization, there is another
layer between the host OS and applications. A thin layer (a small
program component), called Hypervisor, is placed between the host OS
and a guest OS, and applications utilize the services of the guest OS.
A combination of the guest OS and its applications is called virtual
machine (VM) or virtual appliance. This arrangement allows an operator
to host more than one VM on a single machine (physical server.)
Now can this concept be
applied to storage equipment easily? Unlike servers, such equipment is
more complex since there are several types of storages, such as internal disks,
directly attached disks (DAS), network attached disks (NAS), storage
area network (SAN), tape backup and so on. A typical data center has
many of these storage types configured in a complex manner. The
storage virtualization looks at the whole storage areas as one logical
entity and does not distinguish each entity separately. This is a very
useful concept since even hourly or daily, the required data segment (and their amount) needs vary
dynamically due to many factors. By considering the best platform to host them based
on several criteria, such as cost, reliability, quality, accessibility
and speed of access, a particular data segment is placed on the most
appropriate place at a given time transparently.
Compared to server
virtualization, I do not hear much about the actual use of storage
virtualization. However, Tsvetanka Stoyanova of DataCenter Journal,
did some survey work published in his recent article. He says some storage virtualization is utilized in a typical data
center, although it is not widely used like server virtualization. He also points
out that according to experts in this area, storage virtualization is
far from a major technology in data centers now. He
surveyed definitions/classifications and various technologies for
storage virtualization and tried to classify each vendor’s product
according to the survey.
conclusion was there are too many definitions and classifications and
thus, it is necessary to have a consensus on the definition. Rather
than repeating the result, I only point out major players in this
segment as: Dell, IBM, Hitachi Data Systems, Sun, EMC, DataCore,
FalconSor, NetApp, RelData, LSI and leave the details to each reader to
read his blog.
thing is clear. Unlike the server virtualization, the market for the
storage virtualization is still wide open and it is hard to pick a
winner technology or vendor(s) at this time. This market needs to be
Posted By Zen Kishimoto,
Friday, September 5, 2008
Updated: Saturday, September 6, 2008
A new way to develop a data center has been proposed and multiple
leading companies like Microsoft, Sun and IBM have developed and begun
to deploy it. In addition, HP and Dell are rumored to be developing
one. Some details are different among those companies but they are more
or less similar. In order to save several steps to set up and
configure each server and rack at a data center, pre-configured severs
in racks are shipped in a container. A good summary of Microsoft' s
solution is given by James Hamilton's paper at Microsoft Research.
numbers of low-cost, low-reliability commodity components are rapidly
replacing high-quality, mainframe-class systems in data centers. These
commodity clusters are far less expensive than the systems they
replace, but they can bring new administrative costs in addition to
heat and power-density challenges." ……………………………….. "The
proposed solution is to no longer build and ship single systems or even
racks of systems. Instead, we ship macro-modules consisting of a
thousand or more systems. Each module is built in a 20-foot standard
shipping container, configured, and burned in, and is delivered as a
fully operational module with full power and networking in a ready to
run no-service-required package. All that needs to be done upon
delivery is provide power, networking, and chilled water." ............................................... "In
this modified model, the constituent components are never serviced and
the entire module just slowly degrades over time as more and more
systems suffer non-recoverable hardware errors. Even with 50
unrecoverable hardware failures, a 1,000 system module is still
operating with 95% of its original design capacity. The principle
requirement is that software applications implement enough redundancy
so that individual node failures don’t negatively impact overall
Microsoft plans to populate its
Northlake (near Chicago) Data Center, which is under construction, with
220 such containers according to an articlereported by datacenterknowledge.com.
success of the container-based system is yet to be seen. Finally, we
would like to add a note that there are some skeptics on the container
based solution as reported in Eric Lai's article in ComputerWorld.