Volume 5, Issue 39
August 26, 2009
In this issue:
• Two Views - EMC’s New Virtualization-
Aware Storage Solutions
• EMC Launches VMware-Enhanced CLARiiON
Solutions By Charles King, Pund-IT, Inc.
• Navigating Virtualization Storage Waters
with EMC’s Navisphere Manager By David Hill,
• IBM’s POWER7 at Hot Chips 21 — Putting
POWER on the Front Burner By Charles King, Pund-
• VirtenSys Makes the Case for I/O Virtual-
ization By David Hill Mesabi Group
Pund-IT, Inc. Contact:
Hayward, CA Office: 510-383-6767
U.S.A. 94541 Mobile: 510-909-0750
Two Views - EMC’s New Virtualization-Aware Storage So-
EMC Launches VMware-Enhanced CLARiiON Solutions
By Charles King, Pund-IT, Inc.
EMC unveiled new virtualization-aware features for its midrange CLARiiON CX4 networked
storage systems optimized for VMware environments. These include:
• A new version of EMC Navisphere Manager enables virtualization-aware storage manage-
ment. Through integration with VMware vSphere, the resulting solution delivers end-to-end
physical-to-virtual mapping, helping administrators meet Service Level Agreement (SLA)
and Quality of Service (QoS) goals.
• Shared 10Gb/s Ethernet and 8Gb/s Fibre Channel (FC) connectivity options, which allow
CLARiiON owners to flexibly choose technologies with the right blend of availability, con-
nectivity and performance for their storage and business needs.
• New EMC RecoverPoint and EMC Replication Manager’s capabilities which deliver en-
hanced ease-of-use, cost efficiency and data availability for VMware environments. Accord-
ing to EMC, these new features are key enablers for private clouds and ‘fully virtualized’ in-
EMC offers 8Gb/s FC today solutions and will offer 10Gb/s Ethernet on CLARiiON CX4 net-
worked storage systems in the third quarter of 2009. Both 8Gb/s and 10Gb/s Ethernet are
available today on EMC Celerra systems. Virtualization-aware Navisphere Manager and the
new versions of RecoverPoint and Replication Manager are available worldwide from EMC,
its Velocity2 Program Velocity2 Authorized Services Network (ASN) Partners and are also
being offered by Dell.
New EMC solutions expand midrange storage options for VMware customers.
Server virtualization, particularly in x86/64-based systems, continues to drive fundamental
datacenter transformation in businesses of every sort. In particular, solutions, such as
VMware’s ESX Server enable companies to efficiently execute significant, often massive
datacenter consolidation efforts, reducing the total number of servers they require to a
fraction of what was once necessary.
This provides obvious benefits related to acquisition (or CAPEX) costs but also results in
often profound, long-term operations (OPEX) savings. It is also one of the reasons that
VMware has grown from a niche technology specialist to an innovative vendor with nearly
$2 billion in annual revenues, and why Microsoft, Citrix and others have enthusiastically
joined the hunt.
However, as these transformational efforts ripple into virtually every corner of the datacen-
ter, their immediate benefits tend to obscure often considerable challenges. Consolidating
applications and workloads means that an organization requires fewer servers than it once
Navigating Virtualization Storage Waters with EMC’s Navisphere
By David Hill, Mesabi Group
Server virtualization continues to be near the top of the IT top priorities hit list. A recent
Goldman Sachs survey of IT spending priorities showed server virtualization in third place
followed by closely related server consolidation in fourth place. And both are closely corre-
lated with the #1 (and not at all surprising) IT spending priority of cost reduction.
But new technologies can also introduce new challenges and complexities, and so it is with
server virtualization’s relationship to storage. With direct attached storage (DAS), the horse
and carriage relationship between servers (CPU horsepower) and storage (data carriage)
was clear. Both the physical server and its storage were tightly coupled in a one-to-one rela-
tionship (each server and the storage that it “owned”). Even though the introduction of stor-
age area networks (SANs) physically decoupled servers from storage, still each server was
physically assigned an allocated piece of the SAN storage infrastructure. So, the relation-
ship between a server and storage was clear even if the assignment of storage could be
changed from time to time (which was not possible with DAS). Thus, the distinction be-
tween a server as a physical device versus a server as an operating system and the software
environment that it controlled was irrelevant.
That situation has changed fundamentally with server virtualization with the result that
storage has relationships with virtual machines (VMs) instead of physical servers. And VMs
themselves are not tied to a particular physical server but can move around as needed,
such as for load balancing and performance reasons. That capability is powerful and very
useful for the management of workloads and application performance. However, at the
same time, storage administrators now face challenges in gaining visibility into which VMs
are consuming storage resources. This is necessary to ensure that VMs have the storage
they need to run effectively, to ensure that the overall performance of the server-storage
environment meets service level agreements and to be able to resolve any problems that
Quite clearly, without the proper software management tools, achieving those goals can be
very challenging. The problem up until now has been that performing virtual infrastructure
discovery and reporting necessary for management of the virtual environment, including
storage, was very difficult. And that meant that managing the virtual environment of VMs
working with storage was difficult, at best.
EMC's introduction of new automated virtualization-aware Navisphere Manager capabilities
for CLARiiON CX4 has changed all that. The company cited an example of a medium-sized
VMware environment of 100 VMs spread across 10 ESX servers in which the storage admin-
istrator had to view 200 screens for the 10 ESX servers and then record the information on
each screen manually to a spreadsheet! Naturally that required hours to complete, but to
further compound the problem, the process needed to be repeated to reflect any change to
the environment. Obviously, this was a task that overburdened administrators did not look
forward to and one that cried out for automation not only to free up time, but also to re-
duce the possibility for error. But a proper solution should also not inhibit VMware admin-
istrators from moving VMs as needed without having to worry about recording and man-
IBM’s POWER7 at Hot Chips 21 — Putting POWER on the
By Charles King, Pund-IT, Inc.
At the Hot Chips 21 conference in Palo Alto, IBM discussed details of its upcoming POWER7
microprocessors, including advances in balanced design and innovative new eDRAM tech-
nologies, which support dramatic boosts in both chip and system performance and energy
efficiency. Hot Chips 21 occurs a week after a POWER7-focused analyst event IBM hosted at
its Power Systems facility and lab in Austin, Texas, where the company discussed its plans
for Power System solutions based on the new chips. So is POWER7 really such a big deal?
Absolutely, for both technological and market strategy reasons. Regarding the former,
IBM’s mixture of utterly recognizable and entirely new features makes POWER7 one of the
most intriguing enterprise-class processors to come down the pike since, well, POWER6.
Like many of its competitors, IBM is leveraging proven 45nm technologies in POWER7. Like
others, IBM is delving deeper into multi-core options—various systems will be available
with 4-, 6- and 8-core chips, with each core supporting 4 threads. Balanced POWER7-based
systems can scale up to 32 sockets, 256 cores and 1024 threads.
Practically speaking, this means that POWER7-based commercial solutions are likely to
maintain or extend IBM’s continuing leadership in system and RAS (reliability, availability
and security) performance. In addition, Customers leveraging IBM’s Power Virtualization
can deploy over 1000 mobile and flexible virtual machines on a similarly tricked-our 32
core system, providing ammunition for the company to pitch highly scalable Power7 solu-
tions for server and datacenter consolidation projects. Finally, POWER7 will serve as the ba-
sis of IBM’s efforts in DARPA’s next-generation PERCS (Productive, Easy-to-use, Reliable
Computing System) initiative.
One of the most innovative new POWER7 features is an on-chip memory technology IBM
calls eDRAM, which provides 32MB of internal L3 cache, delivering 6:1 latency improve-
ments over traditional external L3 technologies and a 2X boost in bandwidth. IBM’s eDRAM
is not quite as fast natively as SRAM but it requires just one-third the space of conventional
6T SRAM and uses one-fifth the standby power. Bottom line? New eDRAM and other im-
provements mean that POWER7 cores will deliver higher performance at lower frequency
than POWER6 cores and will provide more than quadruple the chip level performance of
POWER6 in the same energy envelope.
What will this mean in the datacenter? According to IBM, POWER7-based servers will en-
hance aggregate throughput by 2-5X and improve overall performance by 2-3X in the same
energy envelope as comparable POWER6-based systems. Despite these leaps, the new sys-
tems will be extremely familiar to IBM clients. POWER7 servers will be available in a portfo-
lio of modular, scalable systems similar to POWER6 and binary compatibility between POW-
ER7 and POWER6 should ease customer transitions. IBM is further simplifying the transition
process by offering customers who buy new Power 595 systems processor book upgrades
to POWER7 and is planning a transition program for owners of Power 570 systems, as well.
This all sounds great, but how will POWER7’s improvements translate in the larger IT mar-
ketplace? Overall, we believe there is both excellent and somewhat cautionary news for
VirtenSys Makes the Case for I/O Virtualization
By David Hill Mesabi Group
When the word “virtualization” is used, the implicit assumption is that the speaker or writer
is talking about “server virtualization.” But other types of virtualization exist, such as stor-
age virtualization. And now I/O (input/output) virtualization is starting to encroach upon
Just what is I/O virtualization? Answering that question requires a brief review of the role
of I/O in a system. The job of a computer is to perform "data processing." A CPU processes
data from its associated transient memory, but the data typically has to be acquired from
storage somewhere, and a network is used to link the two. The movement of data is all
about I/Os (inputs are reads from storage to a server and outputs are writes to storage).
To be clear, the use of a network may be grandiose when using direct attached storage
where a bus or I/O interconnect is sufficient rather than a LAN or SAN, but logically any
connectivity is still part of a network.
Now if I/O needs to get in or out of the server chassis box in conjunction with an external
network composed of such things as cables and switches, an I/O adapter is necessary to
connect and translate network protocols to the CPU/memory complex of a server. If the
network is a SAN, a Fibre Channel HBA (host bus adapter) may be the I/O adapter or if to a
LAN or WAN an Ethernet NIC (network interface card) may be used. However, the I/O
adapter is located within the server box.
What I/O virtualization does is move these adapters from the server box to a new box
called an I/O virtualization switch box external to the server. VirtenSys is joining the I/O
virtualization party (where NextIO has played for over a year) by providing such a switch.
With the VirtenSys solution, a single PCIe cable connection is made between each physical
server and the I/O virtualization switch. What each server now has is no longer a dedicated
I/O adapter, but rather, virtual I/O connectivity through non-dedicated access to I/O
adapter resources on the I/O virtualization switch.
What does this do and why does it matter? One of the big drivers of virtualization in gen-
eral has been that IT assets have been underutilized, a situation also true of I/O adapters.
VirtenSys says it has found that adapters are only utilized 10% to 15% and their virtualiza-
tion switch can raise this utilization to more than 80%.
That obviously can lead to better asset utilization. Using VirtenSys’ solutions, an IT organi-
zation can defer the need to purchase additional I/O adapters until the demand for I/O in-
creases sufficiently to warrant them. And even when additional purchases are needed,
fewer adapters will be required than in the past. Moreover, some currently redundant I/O
adapters can simply be taken out of service. That reduces the power and concomitant cool-
ing requirements. That is nothing to sneeze at as up to 30% of power is consumed by I/O
cards and switches.
Those are obvious benefits, but there are a couple of other points which, although not as
obvious, can be critical. First, however, note that a VirtenSys switch is non-disruptive in
that it works with the existing infrastructure in data centers. That means no changes to