13 April 2015

Universality and Open Ecosystems for the Age of Softwarization

If I were at NetSoft today, I would have opened our flaship conference this way...

It has been mentioned several time that Software-Defined Network (SDN) and Network Functions Virtualization (NFV) are just different facets of an overall systemic and disruptive transformation, the so-called Softwarization, which is going to impact Telecommunications, ICT and other related ecosystems.

In fact, even if SDN and NFV are well-known paradigms, since a few decades, today they are (probably) becoming exploitable and sustainable, given ultra-broadband penetration and the increasing performance and the down-spiraling costs of ICT. As such, SDN and NFV represent just a part of the Softwarization process which will transform our Society: Cloud, Fog and Edge Computing are part of the same systemic picture. And also Internet of Things.

SDN is not the next "networking paradigm" and it makes no sense considering it independently from NFV as from the other paradigms, such as Cloud, Fog and Edge Computing.

It's Softwarization, simple enough, but highly disruptive, opening the way to Open Ecosystems.

in fact, Softwarization is going to bridge several gaps: one is the line between the applications of Terminals, Consumer Electronics, Smart Things, Machines and the businesses of the Digital Enterprise and even more the processes of Industry 4.0. Terminals, wearable and any robot, machine...in fact will become part of a pervasive, dynamic and highly flexible virtual infrastructure capable of handling complex communications and business processes (including mission critical applications).  

Huge datasets produced by pervasive machines, terminals, sensors, and devices will be moved and handled in real-time: massive and pervasive horizontal infrastructures, based on ultra-low latency links and capable of running powerful compute and storage engines. This will end the age of vertical silos.

Then, future Telecom-ICT infrastructures will look like pervasive, massive, multi-domain Data Centers or “Supercomputers”, with fixed and mobile ultra-low latency links providing the required pervasive connectivity. We can imagine then a sort of Operating System (OS) managing the hardware and software resources and providing sets of services, programmable with APIs. On top of this OS, applications will flourish to create and develop the new ecosystems for the Digital Society and Economy. ON.OS is not that far from this model…which should be extended to mobile ecosystems.

The value is moving from hardware to software and the software will be the mean for implementing flexible architectures-of-architectures based on cognitive, deep-learning algorithms, A.I. methods...and other mathematics crunching datasets to infer decisions, to learn and to actuate actions onto the reality.

Main systemic characteristics should be universality (e.g., based on recursivness of a limited number of functional component and blocks) and openness (e.g. adopting Open Source Software and Hardware). Main technical challenges include:
  • What Abstractions ?
  • Which virtualization applying to the different types of resources: e.g., virtualized network functions can be implemented as a full VM using a server node, virtualization based on hypervisor (e.g. QEMU, Xen, KVM, and VMware), as a Linux Container, (e.g. LXC)...
  • How designing an overarching and distributed Operating System capable of orchestrating applications live-cycles, software and hardware infrastructure’s resources ?
  • Which controllers for different kind resources (up to terminals) ?
  • Which levels of “Programmability” through dynamic APIs ?
  • How automating processes: eTOM vs approches a la DevOps ?
  • How “mitigating” CAP Theorem limitations ? The CAP theorem states that any networked shared-data system can have at most two of following three properties: 1) Consistency (C) equivalent to having a single up-to-date copy of the data; 2) high Availability (A) of that data (for updates); and 3) tolerance to network Partitions (P). Looking at a future software-defined infrastructure as a distributed Data Centers or super-computational system (routing/forwarding packets can be seen as a computational problem), just two of the three CAP properties will be possible at the same time.
  • How achieving end-to-end latency of units of milliseconds? The key factor of success…
  • Multi-domain, multi-vendor interoperability for virtual platforms (mind the clash Standardization traditional processes vs Standards-de-facto);
  • How handling Security and Privacy issues “by design” ? We don’t need reinventing the wheel any time ! Let’s consider rule-based filtering…
  • etc…

Eventually, we'realizing that technology might be indisputable but economics and policy rule, so: which are the business models for the newly enabled service paradigms beyond “commoditization” ? which are the policies ? Technology, economics and policy rules will be highly intertwined in the Age of Softwartization!

I hope that NetSoft will contribute bringing some answers to these and other relevant questions. 

Drop me an email if you wish joining our Community !

No comments:

Post a Comment