CENX 執行長 - Ed Ogonek 提到當大家全力衝刺建構物聯網時代，聚焦於更厲害的感測器、數據應用或是終端體驗時，事實上深藏於物聯網之中更重要的議題便在於"網絡"，所有智慧化的過程都在網路、軟體中進行，因此在物聯網的趨勢下，我們更應該注意"網路"這個隱藏的議題。
快看看以下文章，Ed Ogonek執行長怎麼解決?
While everyone focuses on the cool wristbands, industrial sensors,
data-driven applications and the compelling end-user experience, the
real smarts in the Internet of Things (IoT) is largely hidden from the
public, and from the media too: The intelligence is in the network, and
in the software that keeps that network humming.
Make no mistake: The sensors and devices being designed and used in
IoT applications are amazing. Whether geared for consumers, smart
cities, the enterprise or industry, IoT cloud-based applications are
incredibly powerful, and offer the potential to make a real difference
in peoples’ lives. That’s especially true when the back-end applications
leverage Big Data for analytics, and apply machine learning and
automation to truly create services and products that go far beyond
traditional software.
There’s a soft boundary between the IoT and other applications that
link remote end-points with back-end servers. Perhaps there’s no
boundary at all.
My own home is rapidly becoming a smart, efficient home. My company’s
offices are becoming smart, efficient offices. My city is following
suit. Now, let’s talk about the hidden challenge of the IoT: Connecting
all the mobile and fixed devices, including sensors, to back-end
services, whether in the cloud or a service provider’s data center.
If the links aren’t reliable and cannot handle the vastly varying
data bandwidth demands, those billions of devices can’t talk to back-end
applications. Telemetry can’t be gathered from remote sensors. That
means actions can’t be pushed out to remote devices, like
remotely-operated surgical devices or automated sprinkler systems. Big
Data-centered applications won’t have the data they need to make
real-time decisions and carry out real-time actions.
Focus on Commnications Service Providers
The IoT relies upon a vast array of networking technologies, as well as
network service providers. There may be wired Ethernet-based Local Area
Networks (LANs) within buildings, or Wide Area Networks (WANs) between
buildings. There may be WiFi within corporate buildings or public
spaces, like an urban park, and that WiFi might be managed by the city
or hosted by a communications service provider (CSP).
As organizations look at the IoT, they must consider the reliability
and scalability of all the networks. Capacity planning is key to
ensuring that network infrastructure is robust enough to handle
increased data traffic – but that’s not enough. Most large organizations
have extensive contracts with Tier 1 or Tier 2 CSPs to provide WAN
services to link their sites, and also connect their networks to
cloud-based Software-as-a-Service (SaaS).
Enterprises, industries, governments and other organizations looking
to roll out IoT at scale must work with CSPs to make sure their systems
are also sufficiently robust. Contracts should be studied, in particular
for service-level guarantees regarding downtime, bandwidth, elasticity,
latency, jitter, and delay. Many new IoT applications are much more
sensitive to network performance, and have significantly varying service
requirements than traditional applications – especially when so much
traffic is going to and from IoT devices on the network edge.
Smart Service Provider Networks
To get ahead of the curve for IoT services, and to differentiate
their service offerings, CSPs are proactively changing how they manage
and build out their own networks. After all, while some IoT projects
will start at an organization’s C-level and work down; many others will
begin as quiet, unassuming initiatives driven by various
line-of-business departments. In fact, some business managers may be
piloting IoT projects and not even realize it, or understand the full
benefits and limitations of data networks.
To accommodate the increasing demand for IoT services, CSPs are
beginning to instrument their own networks to be more intelligent, more
robust, and more data-driven. Given that IoT projects span multiple
networks, orchestration of intra-network and inter-network services is a
must, in order to assure the basic connectivity and bandwidth required
to provide an array of IoT services. Increasingly, these networks will
include Software-Defined Networking (SDN) and Network Functions
Virtualization (NFV) infrastructure due to the programmability and
scalability that those technologies provide.
The smartest tool in the intelligent CSP network will be orchestrated
service assurance and management, driven by real-time Big Data network
analytics. Advanced intelligent analytics will help carriers optimize
their hybrid physical and virtual networks to be more elastic – better
able to accommodate unexpected data traffic and to ensure a high service
level quality for all the billions of things, irrespective of what type
of IoT application is being executed.
With orchestrated service assurance, CSPs will also be able to
identify and isolate network faults, and route traffic over alternate
paths until the fault can be repaired. In large networks, after all,
faults are going to occur, whether they are caused by fiber
cuts, hardware failure, power outages, configuration errors, or excess
traffic degrading real-time performance.
Finally, orchestrated service assurance can help service providers
meet their customers’ service level guarantees, while simplifying
end-to-end management across all the networks used by the IoT, including
cellular 3G/4G/5G, WiFi, LANs, and Carrier Ethernet WANs.. When the IoT
application needs bandwidth, coverage or specific levels of network
performance, it simply must be there.
Thanks to technologies like real-time Big Data network analytics, and
orchestrated service assurance and management, communications service
providers are the real heroes, the real enablers of the IoT. Those
technologies are making a difference in peoples’ lives every day, even
if nobody knows it.
資料來源: http://www.iotevolutionworld.com/iot/articles/421843-looking-real-smarts-the-iot-hint-start-with.htm
