
Virtualization and Cloud executives share their predictions for 2016. Read them in this 8th Annual VMblog.com series exclusive.
Contributed by John Dickey, CTO & Co-founder, Talari
In a Years' Time, It’ll Be About Time Driving Smart SD-WAN Adoption
The typical
formula for this type of article is to take an inventory of technologies that
are at the prime of their respective adoptive hype cycles
and proceed to wax poetically on how they will finally go main stream in the
year to come. It is good practice to cast a wide net of predictions, so that in
subsequent years, you may be able to note some of the fish you caught while
resisting the temptations to neglect ones that got away. Conveniently for
me, this is my inaugural blog for VMblog, so I have the good fortune of a clean
slate to work with. That said, it is indeed an excellent practice to assess
what has transpired in the pursuit of pondering what is to come. History
trends, cycles and repeats. The hill we are on helps us see where we are and
where we are going. By the fact that you're reading these words means that you
understand that our industry is dynamic. In the last year we have seen dramatic
shaping events that are driving us to new approaches. Here we assess some of
those challenges and contemplate the potential repercussions they present us in
the year to come.
At this time
security is a constant and ever pervasive driver for us all. Companies and
careers can be destroyed in a day without proper care and attention to the
safety of our data in flight or at rest. It is on all our minds, keeping
us up at night, and has compelled us to mandatory action. Safeguarding our data
has driven us to adopt new models that previously we only explored. We are
pulling services back from the outer reaches of the remote branch offices into
to our core data centers where we are able to provide greater protection. This has
been facilitated by the availability of inexpensive bandwidth to our various
locations. Interestingly, many of us are starting to recognize the greater
competence for security may be provided by our cloud partners. The general
thinking goes, nobody has more to lose than Amazon, Google, Microsoft and IBM
and they take great pains to protect their data, so why should our firms not
leverage that hard-earned and proven expertise for our business?
Another
challenge is the ever-growing expectations of the end users for improved quality
of experience (QoE) accessing their tools for their livelihoods. The end users
require crisp response times so that they can maintain productivity. They are justified
in being frustrated when the infrastructure and apps do not
cooperate. After all, you can only be as productive as your VDI permits
you to be. Additionally, accelerating this trend is the growing need for
anytime, anywhere peer-to-peer collaboration. Unified communication, web
meeting and web mail all are sensitive to network latency and unforgiving when
the network under performs.
In addition,
the end users and the business are driving us for greater flexibility in how
they work and where they work. Yes, they BYOD, but also they desire a BYON
(Bring Your Own Network). Additionally, with growing frequency and with
the company credit card in hand, they are bringing their own SaaS apps. We
understand the end users and the business units they serve are having to be
more productive and adaptive to compete in a challenging interconnected
world. We understand that they have to take that call from Hong Kong at the
hotel in the middle of the night. They need the communication, they need the
app, they need the data and they need our help. We are trying but it is
not easy. We know for the end users in particular, there are less and less
obvious network walls for us to firewall. We understand that any
unencrypted data in flight, even when on premise, is going to be no longer
acceptable. Security is mandatory but we have to provide for it in ways
that allows for flexibility and productivity as well.
Thus, as the
data migrates to more secure distant and secluded locations in the core or
cloud, the share of critical time-sensitive traffic is growing at the periphery
driving the demand for higher QoE service levels. You may be pondering WAN
optimization as a means to resolve this, but regretfully, its value is
diminished by exactly these same trends. WAN optimization does not help
with interactive and real-time traffic. For bulk encrypted traffic, you cannot
data reduce what you cannot not see. There are administrative workarounds
for the encryption issue. For example, you could try and seed keys at the
appliances to permit decryption and stashing but it has logistics issues and
may still fail to satisfy. Even if you overcome the administrative challenges,
the net effect is to have a man in the middle appliance storing data at rest in
the less secure remote location, which is contradictory to the goal of
centralized security. In addition, in 2016 because of continued evolution
of WAN-friendly, end-to-end applications over plentiful inexpensive bandwidth,
the WAN optimization market will accelerate in its decline. Indeed,
Microsoft has fixed the glitch and your data is too valuable to risk it. It is
notable in 2015 how many traditional WAN optimization companies pivoted to Software
Defined WAN (SD-WAN) approaches because of these issues and it is well
justified. How they fare in the new world of SD-WAN is yet to be seen.
Other
traditional approaches use static proximity of services to manage the
challenge, such as CDN, but again this has limitation. If your data and
applications are distributed around the world and your users are mobile and
dynamic you are constantly in a "whack the mole" mode of building and
maintaining static infrastructure to chase your dynamic business needs. So the
old approaches no longer satisfy but nonetheless, we have the new challenges of
time-sensitive and often unpredictable traffic at the periphery UC
collaborating while simultaneously interacting with the secure data in the
core/cloud.
It is our
belief that 2016 will bring much greater adoption of unidirectional bandwidth
latency management methods provided by Smart SD-WAN solutions. For
Talari's technology, in particular, these algorithms are as fundamental to us
as search is to Google. Smart SD-WAN solutions have unidirectional, multi-level
instrumentation and continuous real-time analyses that enables automatic and
seamless adjustments on a sub-second basis to packets traversing the network
assuring end-user QoE. The latency analytical techniques are not new and
have shown their utility for over a half a decade at thousands of locations for
demanding mission-critical solutions. In 2016, what is changing is the
accelerating need for it in vastly more enterprises driven by the
aforementioned rapid changing use cases. In 2016, Smart latency aware SD-WAN
autopilots will step in to fill the QoE gap. They are dynamic, agile, industry
proven, ready and able. In the time to soon follow these same key metrics will
be utilized to drive the next generation of automated and integrated end-to-end
SDx and SFV service chains.
So our
prediction is that in a years' time, it'll be about time driving Smart SD-WAN adoption.
##
About the Author
John
Dickey is the CTO and co-founder at Talari. Before co-founding Talari, John was
senior director of software engineering at Applied Microcircuits (AMCC), where
he led the architecture, design, and development of a broad array of switching
and routing protocol software implementations for embedded systems. John joined
AMCC upon its acquisition of MMC Networks, the pioneering maker of
network-optimized processors (NPUs). He was also the founding Chair of the
Network Processing Forum industry standards organization's Software Working
Group and holds patents in multiple areas of networking technology. Prior to
MMC, John held a variety of engineering and management positions at IBM and was
distinguished as an Advisory Engineer. John earned a degree cumlaude in
computer science from the University of Pittsburgh.