Sheikh is Huawei Middle East Senior Architect for NFV , Telco Cloud ,SDN with focus on ICT Service delivery through Telco DevOps . Focussed to define the Roads for future 5G Core Network . Always interested in those disruptive technology driving the industry transformation
What are Telco Applications in software means ? . A modular ,segmented code with defined API’s that can be customized for different functions and requirements .For Telco applications with varying requirements including uneven usage, bandwidth ,spikes during periods, the requirement for having built in elasticity and scalability is crucial for larger POD’s environments .However for such cases the applications should be designed to detect variations in the real-time demand for resources including but not limited to bandwidth, storage, compute , SLA and KPI to add for Telco use case. However, in legacy world applications had been developed to run on a single machine and require re-coding to adapt for both the scalability and elasticity that the Telco cloud provides.
The scenario that Telco Applications are persistent in nature make the scnerio more intricate to achieve
There are two main concepts to understand between Elasticity and Scalability in Telco Cloud.
- Scaling Horizontally or Scale Out /Scale In:
Those guys who have delivered a NFV based Telco Cloud knows the Scaling is supported currently , what it actually means is that EMS will notify current load to VNFM in real time or Polling way and it can make decision to add/remove a new VM and launch new instances of VM for the application . This do require an algorithmic policing to balance the load of vAPP on all new and old VM’s in a controlled fashion. Normally this is practice used in NFV Projects now
- Scaling vertically or Scale Up /Scale down
If you are not following industry may be you will not be well aware what this will mean for NFV? This is concept normally coming from AWS use case of Netflix for Persistent workloads a use case which AWS supports for large Enterprises. It will mean that moving a vAPP real time to a bigger /smaller VM or resizing the VM.
- The myth of Scaling
As we can see as long as the vertical scaling in Telco environment can have a Telco DevOps as I use to call the process to verify Build, I&V, Test the service in somehow automatic way .It will make a perfect case to avoid trap of load balancing and validation through long exercise of manual testing and MOP following by front line.
Incorporating this model through OPNFV test projects like Functest and Yard stick will mean this model can help for fast scaling and complete DC expansion more quickly. But obviously Industry need to consider how to assure no impact on VNF performance, SLA, KPI and that such scaling must be real time on persistent workloads?
Currently this case is not supported???? may be one alternative is use Telco service capabilities to isolate certain VNF first and then apply this model, but for sure it will have challenges in case VNF is split across sites in a federated network?
This model will be the key to deliver agility on big Telco DC , the model 1 as used now in NFV world will be obsolete in coming 2-3years after all Telco’s also want to become the Netflix of industry ………………………….:)
It will mean NFV world will move one step ahead from Being Scalar to being elastic for the workloads.
- Key Challenges
However to evolve to this model requires Industry to agree and solve some key issues
- How to solve FCAPS and concerns from operations customer
- How to assure KPI before and after
- 100% automated test after expansion
- 100% automated performance validation
- How to introduce service for workloads? Consider there will be running traffic before expansion
- How to address customer experience for such cases
- Watch the Road Ahead
Finally to wrap us we must also think how soon Openstack the unified standard for Telco NFV will support evolution from PNF to VNF and subsequently Kubernetes can deploy on it to support Scaling through containers . It is sure finally this will solve issues but till the Telco world key concerns as higlhighted above are not adressed . The road to reach there seems gloomy . Do watch below video to see whats happening latest in this domain as a take away from Boston summit .
I will like to know your thoughts on this matter , because for persistent workloads the scaling have more challenges than ephemeral workloads and journey is challenging but interesting .
This will be the key to success !!!
I am a Senior Architect with a passion to architect and deliver solutions addressing business adoption of the Cloud and Automation/Orchestration covering both Telco and IT Applications industry.
My work in carrier Digital transformation involve Architecting and deploying Platforms for both Telco and IT Applications including Clouds both Open stack and container platforms, carrier grade NFV ,SDN and Infra Networking , DevOps CI/CD , Orchestration both NFVO and E2E SO , Edge and 5G platforms for both Consumer and Enterprise business. On DevOps side i am deeply interested in TaaS platforms and journey towards unified clouds including transition strategy for successful migration to the Cloud
Please write to me on firstname.lastname@example.org