The cloud adoption is now trivial to almost all
business verticals, the trends suggest organizations, OEMs and developer
communities are looking way beyond cloud in next five years. In the coming
months and years, we will see massive reorganization of the "Cloud", the future of cloud will be an augmentation of the
technological breakthrough that challenges the very assumptions around the
cloud stack and the cloud native applications, devices and technologies truly a
transformation of the cloud enabled by cloud!
We can safely see the trends of the cloud
transformation unwinding on few major fronts..
as a Service,
of data / storage requirements,
perimeters with IoT and 5G..
buzzing AI/ML that
is driving almost all technology ecosystems to our augmented intelligence of
cloud, so to say!
So, What's the
Any cloud provides on demand resources on compute,
storage and memory coupled with interconnected network that ties the resources
together, what is current missing is the standard design principle that can tie
multiple technologies and vendors together creating a single nimble cohesive "Cloud"... although the applications are getting on the
bus with RestAPIs and SaaS models, but the core elements of cloud orchestrators
are still native to the cloud thus binding the customer workloads to the
orchestrators and abstraction layers that twist the technologies making them
highly adverse on interoperability, the migration of one workload from one
public cloud to another is difficult and has to undergo interventions and
conversions due to lack of interoperability becoming a chore.
There is thus a market segment getting created that
just helping to solve this problem and does journaling ( Read,
change logs) and migrations between public clouds and hypervisor replication
needs thereof. This needs to change for seamless workload migrations and
replications since loads of old applications still need the DR via replication
and aspire to become "cloud-native".
What does the 'Cloud Native' mean?
"Cloud Native" is an approach of
designing, building and running applications, services and eco system of
continuous integration, container engines and orchestrators, modern databases
etc on top of infrastructure-as-a-service offerings of cloud service providers
with an objective is to deliver the services with speed ( read reduced time to
market), scalability (Read, Infinite resources on demand / run time) and
finally achieve cost benefit.
- Scale: as business grows, it becomes strategically
necessary to support more users, more locations, augmenting capacity on the fly
with a range of devices, while maintaining responsiveness, managing costs, and
most importantly availability.
Benefit: in this world of
infrastructure-as-a-service, a strategic goal may be to pay for additional
resources only as they're needed - as new customers come
online, more and more transactions are happening point in time. Spending moves
from up-front payment i.e. CAPEX to OPEX (as operating expenses). But this is
not all, being Cloud Native is usually to spend less on hosting than 'owning' the resources.
- Speed: companies of all sizes now see strategic
advantage in being able to move quickly and get ideas to market fast leveraging
Cloud. it is about managing risk of time to market if the entire hardware stack
was to be owned and built on your own. At its best a Cloud Native approach is
about de-risking as well as accelerating change, thus allowing companies to
delegate more aggressively and thus become more responsive to their business
demands and markets they operate within.
Typically, Cloud Native promotes five architectural
- Use infrastructure-as-a-service: run on servers that can be flexibly provisioned on
demand is the essence of this principle, in today's
dynamic world of changing business priorities and evolving business models
making upfront investments on the technology stack and expecting a return over
years is not viable option for many organizations, the "As-a-service" model is fast catching up that not only reduces the
impact on fiscal objectives but also reduces the chores of managing the in
house data center, technology refresh and AMCs on a longer period, the
scalability and on demand consumption make this option as viable as it can get.
- Design systems using, or evolve them towards, a
individual components are small and decoupled from the monolithic large legacy
applications thus requires lesser compute yet supports large userbase or
transactions / event handling in the same set of hardware, the reengineered
applications now has ability to scale and deploy in real time and scale
automatically based on the demand.
- Automate and encode: replace manual tasks with scripts or code. DevOps
could be the right word to describe the automation of all the tasks that can
and should be. This is a productivity leverage over and above other principles
- Containerize: the granular approach to package processes with
their dependencies making them easy to test, move and deploy in run time and
- Orchestrate: abstraction of hardware resources and pooling them
to be presented to the workloads creating demand in production using
off-the-shelf management and orchestration tools ecosystem.
While entire world is going after making their application
and data "cloud-native" there are subtle trends
that are transforming the cloud itself, let us touch upon few emerging themes
that are now trending and we can see their footprints in the applications, data
and orchestration engineering efforts by multiple providers. I tried
classifying them largely into these Six buckets to make it easy
for our understanding and putting forth the context.
- Dynamic Abstraction Separating Applications, Hardware and Data, runtime
configuration and enablement of resources, there three tenets are connection,
consumption and capabilities of the cloud...
- Connection - the smart devices are claiming their share of the
pie, cars, streets, sensors, wearables, thus experiences are getting smarter
thanks to IoT, analytics and AI/ML engines all focusing on a better and real
time user experience driven by events that in-turn changing the very nature of
transaction driven system of records. The devices are connecting to cloud and
changing the dynamics of the so-called application driven transactions, the
three tier, web tier is legacy, the distributed and event driven interactions.
- Consumption - the on demand, pay per user and outcome-based
resource subscriptions is redefining the IaaS, PaaS and SaaS models into
fractions of the experiences typically lead by MobileApps, wearable sensors,
smart devices that are driving consumption-based services rather than subscription-based
services. The cloud is not limited to storage, compute anymore the code as a
service running as Serverless is a new channel that is mobilized by the cloud..
- Capabilities - the cross stitching of elements that create a
dynamic workflow via pooling various resources and run the tiny workloads on
containers, codes from on-device, edge or at the core of "Cloud" demand capabilities of resource pools that wakes up,
runs the workloads, interoperate and kills itself once the session gets over,
truly event driven computing. When the purpose of computing getting serverless
and expectations are building up strong about the underlying infrastructure
layer remains consistent and available in any adverse condition duly served by
different stack or location keeping intact the user experience.
- Faster, cheaper and Distributed processing - the evolution of compute is now getting tied up to
the evolving data, application and AI/ML architectures that demand for parallel
processing for distributed applications, modern databases too have dictated the
need. The processor cousins, CPU, GPU, TPU and FPGAs must now coexist and
create "processing on demand"
ecosystems for modern workloads. The processing is no longer remains a silo and
newer ASIC ( Read, Application Specific Integrated Circuits) chips are
integrated with deeper integrated application functionalities for serving the
purpose to stay ahead of curve, duplication of cores, higher clocking etc are
old school stories, few modern processors now vouch for deep learning, neural
network processing ready and bridging the gaps between the CPU / GPU terrains
of yester years... there is lot happening there!
- Distributed Modern Databases and data stores - the modern database systems are mostly open source
and already have gigantic testimonials for scalability and availability by
design. Gone are the days where databases were supposed to follow sacrosanct
procedures just to keep them up and running, tunes and backed up in case of
failures, the bigdata revolution kicked in the CAP ready - Consistent, Always
Available and Partition tolerant databases extending the data layer from
devices, edge and distributed data centers to multiple cloud orchestration
layers. They seem to be aware of the dynamic scalability and distributed
paradigms of the container / serverless technologies.
- Any Cloud Compute - Fundamentally the change in the way resources are now
consumed at the edge to the core, the public, private, hybrid clouds must
seamlessly interconnect creating cloud as a resource layer that seamlessly
extends the perimeter collaborating with different public clouds, orchestrators
into a single cohesive governance structure managing varied pool of resources
that can will be deployed in a single click with granular governance of
policies for security tagged to the workloads running in the instance of time.
This itself is an area where multiple startups are pitching in to transforming
- Imagine once we have all above charters in place and
the cloud start acting global, the onus will still remain
on the interconnected networks with SDN / SDWAN capabilities from within
datacenter leaf and spine. WIFI networks to 2/3/4G and emerging 5G over the air
networks that will ultimately bring devices, datacenter and data together. The
rural penetration, fiber networks even the power line networks will play a major
part in the future for transforming the seamless experience user demands
despite his location and proximity to urban centers where network availability
and reach is obvious and taken for granted, just imagine a 100 acre
agricultural field un-manned by the devices, robots and sensors strive to
remain connected with each other and the mothership algorithms running in the
cloud, they all will take the share of bandwidth and speak multi-protocol
multi-casts.. the future thus will actually rest on,
always on, always connected state. Just this week, the release of Draft National
Digital Communications Policy, 2018 now awaits the cabinet approval is promising 5G to
50Mbps broadband connectivity taking India to the next league of
- Security and Governance - with recent GDPR enforcement Information security
remains the concern for the board, the penalties tied up with the global
turnover of the company now draws clearer accountabilities and larger impact on
the brand. Less said the better on the awareness and inclination of several
large organizations yet to get their acts right on the information security and
respecting the obligations on both legal and moral grounds, the new AI tools
that are being engaged by the hackers are already seeping into the threat maps.
While Srikrishna Committee submitted its initial assessment and recommendations
on data privacy and management last week in a 176-page report, as well a draft of the legislation on data
protection titled Personal Data
Protection Bill, 2018. Even
the new draft of personal data protection
bill framework inviting the
public feedback is worth a look, the whitepaper can be referred here. Interesting to see the
developments in the upcoming monsoon session of the Indian Parliament for
further discussion on this bill advocating for stringent data privacy and
protection. In one of the judgements, The Supreme Court of India stated that
the "right to privacy is protected as an intrinsic part of
the right to life and personal liberty under Article 21 of the Constitution and
as a part of the freedoms guaranteed by Part III of the Constitution". Given the public and the private sector are
collecting and using personal data at an unprecedented scale and for
multifarious purposes. While data can be put to beneficial use, the unregulated
and arbitrary use of data, especially personal data, has raised concerns
regarding the privacy and autonomy of an individual is thus becomes matter of
concern. With rapid ingestion of technology and inferences caused the
information security is becoming core issue of the digital transformation and
the notion of being in the cloud will continue to raise eyebrows. CSPs and OEMs
continue to remain vigilant about the hardening and whitelisting of their
technologies and tools via product and services offered whereas cloud consumers
keep tab on the people and process aspects balancing best of both worlds, I may
In Conclusion, the journey to, within the cloud
and beyond remains interesting continues pushing stakeholders on a higher
learning curve. Digital transformations, technologies (Read, AI/ML, VR/AR, IoT
and BigData) driving industry reforms and governance are the key tenets for Cloud
unleashing endless opportunities for the future of IT.
For the first time we are able to see government
talking about National Fiber Authority and National Digital Grid towards
digital transformation while the business models too are transforming as
technology feels tangible, lucrative and within reach of a common man as the
computer are now beginning to see, listen, talk, relate and decide within the
rationale set forth by its master, the margining lines between the business and
technology leaders are evident and may finally evolve into the rainbow, a
promising sign of the future indeed!