March 8, 2017

343 Views

How To Realize The Full Value Of Software-Defined Everything?

Co-Authored by : Sachin Mudholkar

DEFINITION OF “SOFTWARE DEFINED EVERYTHING

“Software defined everything” is a container term indicating the shift towards a controlled way of accessing network, data center, and storage. It is aimed at reducing silos in the market. The belief is more on open source APIs.

Led by software defined networking, SDE adds more programmability to the IT infrastructure by decoupling the management part from the underlying hardware. It is flexible enough to handle elastic requirements of an organization in terms of user need.

From the era of manual configuration, we transitioned to a script and GUI based infrastructure.

Today, we are striving towards a programmatic way of creating IT infrastructure to handle dynamic user needs

Components of SDE

The evolving software defined based technologies across legacy networking, storage and data center management are drawing attention to meet the needs of diversifying communication services along with higher data traffic and speed demand. 

Reduction of maintenance cost, a boon for organizations

Cost implication with SDN:

Starting from an agile employee account management to setting up a new client network, manual intervention is significantly reduced. SDN offers seamless network connectivity of plug and play devices, paving the way for more IoT/M2M nodes and other infrastructural requirements. With more demanding customers, network downtime is highly unacceptable. Error prone and time consuming manual network configuration needs to be avoided in such circumstances. With SDN, the need to maintain manpower round the clock to predict upcoming issues is minimized. A connected lighting system enables citizens of smart cities to experience a new era of illumination and lighting with lesser manual efforts as well as cost, since the controller is centralized for multiple nodes.

How about SDS?

The prime point is simplified architecture, which requires people with lesser specialized skill for system administration. With the rise of cloud and analytics system, limitless data management is handled with elasticity, where manual supervision is impossible as the need of the hour is not predictable. The mix of software with storage makes it possible to work with heterogeneous and commodity hardware, reducing the total cost of ownership.

Wondering how Facebook’s huge data center of 307,000 sq. ft. is managed?                                        

It is cumbersome, but will cease to be so in the future due to SDDC benefits. Managing a data center will no longer be a challenge, thanks to various orchestration and automation techniques. 

Local interaction with devices leads to less field engineers

With SDE demonstrating great prowess in simplifying business processes and operations, we can foresee that the field engineers’ working scope is going to reduce significantly. So, what next?

On the bright side, the management tasks of engineers will reduce considerably. Hence, they can focus more on innovation than on performing menial jobs.

However, there are certain impediments:

  • Everything can’t be tested with commodity hardware. So, retesting is required once we have something new to interact with.
  • Complexity of the infrastructure plays a crucial role. The number of controllers seeking high maintenance should not be more, while a lesser number welcomes more failure occurrences and security threats. More effort is needed to protect controllers.
  • Software defined data centers are assumed to be carrying risks along with it. Some organizations need to focus on security on priority rather than enjoying the SDE benefits.        

A large and dedicated workforce containing certified and specially trained people is no more needed to perform routine jobs. But skill is still a critical requirement while shifting towards SDE. Since it is not going to be a black and white transition to SDE, the grey zone gives some time to prepare the engineers to understand what’s coming next and to act accordingly. Also software defined “hybrid environment” is more expected that software defined “everything”.

Conclusion 

It is agreed that the major benefits of SDE are cost optimization with hardware as well as manpower reduction. But to enable automation, human effort is still required for designing and implementation.

Before one loses significance, a robust engineering skill should be adapted for a more scripting/ higher level programming language, facilitating efficient system programming instead of just relying on CLI/GUI outputs.

The most important goal should be to understand the fundamentals of the underlying system along with the current trends and program the infrastructure accordingly. The model of an organization can be changed to set an example in the area of resource management. 

References

http://www.computerworld.com/article/2484358/it-careers/will-software-defined-networking-kill-network-engineers--beloved-cli-.html

http://www.networkworld.com/article/3138475/infrastructure/is-infrastructure-relevant-in-a-software-defined-world.html
http://www.datacenterknowledge.com/the-facebook-data-center-faq/