Our 21st customer event was held on 05–06 October at the (by now) familiar
venue of King’s College, Cambridge, UK. An intimate gathering of new and regular
users engaged in animated debate at every opportunity on the two dominant themes
of network virtualisation, as well as further game-changing developments in the STEM platform.
Crucial exit feedback confirmed that the showcase material “really struck
a chord” and was very relevant to current research activities for many participants.
The whole interaction was described as “very well provided for and executed”
and, in at least one case, “absolutely worth the LONG flight!”
If you would like to read about the sessions in detail and access some of the model
files discussed, please download the full proceedings of the event from the panel
to the right of this article. (A one-time registration is required.) Next year’s
event is scheduled for 27–28 September 2017, and at the same venue. Watch
this space for the provisional programme announcement in April and check your inbox
for an invitation in June.
Showcase models on network virtualisation
In the opening session, Robin Bailey (informally known as
Mr STEM) presented a business-model framework for the
value-chain in a data centre which allows for a detailed comparison
of buy/build scenarios for Anything as a Service
(AaaS) providers from niche player up to the web-scale enterprise. The analysis
considers the respective scale, focus, value-add, efficiencies and margins of players
at all levels from the physical elements of the data centre – compute, storage
and networking; racks and cooling; space, power, fire-suppression and batteries
– through to IaaS and PaaS subscription products.
After the first break, the focus switched to a second model exploring operational
tasks for service fulfilment and assurance relating
to a migration to virtual service delivery. Originally conceived as an adjunct to
the data-centre model, the comprehensive analysis was developed separately and more
generically for practical reasons and benefits greatly from the calculation consistency
and transparency provided by the standard collection and time-factor elements in
STEM. Key details included the basis of average connections for driving assurance
tasks and the significant staff efficiencies facilitated by SDN and NFV features
which can be orchestrated from a minimal number of Network
Operations Centre (NOC) sites.
Guest presentations reflected these main themes, with
two attendees speaking on security considerations and data-centre interconnect in
a software-driven world, and one sharing a recent and rapid learning experience
The next steps session on the two virtualisation showcase
models included a healthy debate on the evolving IT and communications value-chain:
- how the data-centre model could be further layered to differentiate between data-centre tenant, facility provider and building landlord, and
- how the ascent of software-defined networking promotes an increasing layering and
range of actors and opportunities in the connected landscape.
As a practical example of business modelling, workshop style, we added a virtual
firewall service suggested by the audience to the data-centre framework; a compelling
showcase for the agility of the STEM platform as well as the applicability of the
value-chain model. We also looked at the specific example of a managed-backup service,
using a simple, standalone business model to illustrate the clear opportunity for
a new entrant to steal a march on current players which price on GB assured rather
than storage-grid consumed.
The tutorial session which opened day two turned into
a philosophical debate on the ‘left-to-right’ or ‘right-to-leftness’
of business modelling, more meaningfully expressed as ‘platform before services’
or ‘services before platform’. There was reasonable consensus that,
even if you build the platform first and then look to monetise it, ultimately it
is the services that consume the platform and which may even drive expansion of
that platform. It is important to see and even represent the value-chain both ways
in order to speak to CIOs and IT buyers whose natural prerogative is to see the
world ‘from the inside out’ in contrast to the ‘outside in’
perspective of the commercial service provider.
STEM platform development
The formal proceedings for day one closed with a top-up training
session on STEM 7.5, the original version of which was quietly released to a small group of early adopters mid-July. The STEM 7.5a platform update (described in a separate article) which coincides with publication of these proceedings actually responds to some feedback received during this session. From the joy of zoom in the Editor to the efficiency
and elegance of the re-vamped graphing interface for results, STEM 7.5a is now a
recommended update for all users.
The session on extreme detail ‘sorted’
was dominated by a breaking development on the concept of an element partition:
how to transform an informal resource collection tree into a guaranteed-complete
cost hierarchy, including drill-down navigation of results, which we will explain
in a future newsletter.
Anticipated by numerous questions, requests and tantalising remarks from the host,
the regular closing future development clinic heralded
two further profound enhancements that stand to significantly reduce the mental
leap required for a new user to grasp and get comfortable with the power that STEM
offers. Again, in a future newsletter we will share more details on how:
- results will be available directly from a selection of icons in the Editor and graphed
as part of the same Editor view (or as separate views, according to your preference)
- the Editor will offer a Trace Results command to
automate the routine checking of key calculation steps which must currently be looked
up and graphed by hand.
With STEM 7.5 barely out of the stable, it is too early to say when these massive
features will be available, but the expectation is that the next major version will
not be a 7.n!