Cisco has launched its new NCS 6008 giant router in Australia
Share on Twitter.
Get the most reliable SMTP service for your business. You wished you got it sooner!
July 21, 2014
Almost a whole year after the hype first hit the blogosphere, Cisco has done its first installation
of its NCS 6008 massive router, at Australian carrier Telstra. And this is important, since the
NCS 6008 has the potential to revolutionize things.
Back in September 2013, Telstra was said to be one of the customers deploying the Cisco NCS, a
statement which in retrospect could be described as optimistic.
Telstra last week blogged that its 1 Tbps-per-card-capable NCS 2008 has gone live on the Telstra
Internet Direct network, which serves fast internet connections to enterprise, government and wholesale
“Our new Cisco super router is currently managing traffic on the busy Melbourne to Sydney path and
makes Telstra the first Telco in the world to enable live network traffic on the next generation of
routing technology,” the carrier said.
With eight cards, the new system provides an aggregate 8 Tbps of throughput. And it is said that the
NCS 6008 consumes 60 percent less power per gigabit of traffic than the core routers it replaces, the
carrier says, and has a smaller footprint.
Its faster update and boot process, Telstra says, will also help reduce downtime. Next for
the upgrade will be international gateway sites, followed by other national-scale routes.
In other IT news
In the last few years, IBM's storage revenues have dropped, with only Flash Systems showing
some growth. In fact, it appears that Flash Systems is the only area where there is some serious action
Within the Systems and Technology segment of IBM's second quarter results for this year, revenues
from System Storage decreased 12 percent, while at the same time, flash storage grew more than 100
Overall, this is the eleventh quarter in a row in which the Systems and Technology business' revenues
So, removing the storage revenues from IBM's current results and plotting them by quarter and year
we get a clearer idea of how each quarter is doing in a succession of years.
A few years ago, IBM went on an acquisition spree to improve its storage product range. That seems
unlikely to be a course of action IBM will take at present. We would suggest that the biggest
strategic threat to IBM in storage, servers, software and services is the public cloud.
If enterprise customers are not buying IBM hardware to run IBM middleware and using IBM Global Services
to produce the systems that link this hardware and software to business processes, then IBM revenues, net income
and EPS will all fall.
Could it be the case that the hybrid cloud is only a temporary stepping stone, in a timescale of decades, to public
cloud computing dominance? Perhaps the only sure way to compete with the public cloud is to have your
own cloud offering?
That's what Big Blue is currently doing with its SoftLayer initiative. Stifel Nicolaus'
Aaron Rakers noted-- "IBM expects to continue rolling out its SoftLayer data centers in the third quarter
and 4th quarter of this year.
IBM announced its intention to invest $1.2 billion to expand its cloud data centers, planning to
open 15 new data centers worldwide in fact, adding to its existing 13 data centers acquired via
SoftLayer and twelve from IBM.
In other IT news
Intel said this morning that is has tweaked its Xeon firmware and sold it off to Oracle, in an
effort to streamline things a bit. The two companies announced late yesterday that they had partnered to work on a Xeon chip whose
cores can be turned on or off and where the clock speed can be easily changed and tailored by Oracle's
To be sure, Intel's new Xeon E7-8895 v2 SKU was carefully 'adjusted' by Intel to allow Oracle's
code to choose how many of the chip's fifteen cores to use at any given time, and at what clock speed,
said Intel spokesperson Pat Buddenbaum.
Although these three new features-– clock rate manipulation with Turbo Boost, selective power C-states,
and the ability to dial-up and dial-down cores have been available in prior Intel chips, this marks the
first time they've all been available in a single Xeon CPU, we're told.
This allows Oracle to run systems at full processing capacity when a heavy workload arrives,
and spin down the CPUs in quiet periods, all of which should save some energy.
Additionally, the fewer cores used, the higher the clock frequency can go in Turboboost mode,
which is useful for certain kinds of processor-intensive applications.
The 22nm-process E7-8895 v2 has fifteen cores running at a typical clock speed of 2.8 GHz, though
it can use "Max Turbo" to scale up to 3.6 Ghz.
It has 37.5 MB of data cache. Its maximum thermal design power (TDP) is 155 Watts, which in technical
terms is a very hot chip when compared to other similar designs.
The CPUs must be able to reliably work over a range of several clock frequencies, grouped into
grades called "bins". One bin represents a 100 MHz increment in clock frequency for the E7-8895 v2,
Intel told us.
"Let's just say that this is a fifteen core part at the outset, running under mostly favorable conditions
that allow for Turboboost. It can theoretically run at one bin or two bins to three bins above rated
frequency," said Intel.
"Depending on all the many conditions that enable that, you can hit any one of those points. Three
bins is the maximum at the 15-core mark, but as you go lower, say to a six-core mode, then the maximum
frequency within that particular situation is 3.4 GHz. If for some reason only one core is being used,
then in that particular scenario you can go up to 3.6 Ghz," said Intel.
The chip giant spent about a year working with Oracle to help the company tweak its software and
schedulers to automatically scale core usage up and down, and make better use of its power management "C-state"
technology that Intel made available through firmware tweaks, Intel said.
"It's more through the unlocking of features in the chip, but also through our manufacturing
process there will be a more stringent thermal profile required such that the chip supports all of
these different modes," Intel added.
This marks a continuation of one of Intel's major new strategies-- customizing its high-margin
server chips for specific customers, like eBay or Twitter.
Another aspect of this is its announcement of sitting FPGAs directly on top of Xeons for off-CPU
Previously, Intel's customizations have involved different power management methods and selecting
some chips for particularly high sustained clock frequencies, but Buddenbaum believes this is the
first time Intel has sold a chip with dynamic core-scaling targeted at servers.
Though Intel makes much of the customization itself, it hasn't done anything from a hardware
perspective besides select for the best Xeons coming off its production line.
The magic is all in the firmware unlocking some chip features. The new chips will sit inside Oracle's
just-announced Exadata Database Machines X4-8, which have been built for Oracle's "in-memory" database
This new server design is "specifically optimized for a new generation of workloads: database as a
service (DBaaS) and database in-memory. With up to 12 terabytes of DRAM memory, the Exadata Database
Machine X4-8 can consolidate hundreds of databases and can run massive databases entirely in-memory,"
The new server can feature 12 TB of memory per 42U rack, 672 TB of disk storage
and up to 44 TB of PCIe-linked flash as well.
"This new customized version of the Intel Xeon processor E7 v2, developed in collaboration with Oracle, helps
maximize the power of the Exadata Database Machine X4-8 by elastically accelerating peak performance
of database operations, while also reducing the data footprint," said general manager of Intel's Data Center
Group, Diane Bryant.
In other IT news
Big Blue has been following the amazing success of Apple's very popular iPhones and iPads
with great interest in the past two years, and now the IT behemoth has decided that it wants
part of the action.
IBM and Apple late yesterday announced a new partnership to maximize each company's strategies.
Starting sometime this fall, Apple is delivering devices exclusively to IBM, which the IT giant will
then integrate with industry-specific apps for business and enterprise customers.
The idea is to provide specialized Apple devices to companies in banking, health care, insurance,
retail, travel and transportation, to name just a few.
Big Blue said it will carefully listen to the specific needs of its huge list of business clients
spread in over 250 different industries.
IBM will then craft tailored software that addresses those distinct needs. "For the first
time ever, we're placing IBM's renowned big data analytics services at Apple iOS users' fingertips,
which opens up a huge market opportunity for Apple," Apple CEO Tim Cook said in a statement, calling
the new partnership "a radical step in the immediate future of both companies."
IBM CEO Ginni Rometty promised a "transformation to the way people work, industries operate and
companies perform over the long haul."
No specific date has been set yet, but beginning sometime this fall, the iPhones and iPads will
only be available through IBM representatives to meet their business clients' distinct needs.
They will include a private app catalog, data security services and a lot more, Apple and
An IBM spokeswoman said it was too early to give some examples of what kinds of business apps would
Get the most dependable SMTP server for your company. You will congratulate yourself!
Share on Twitter.