2017 Road Ahead for the Maintenance Solutions in Converged Clouds and Hyper scale networks

 

We are in 2017 the 2016 was a bumpy year for the NFV/SDN industry realization .Together we have come far to enter from Concept phase to commercial realization phase, it seems the Clouds are mature now and All large Opco’s and small enterprises alike are happy to achieve the real advantages from cloud which is Scalar, High performance and Modular but the story is not over.

The 2017 business model of all large Tier1 Operators is circumvented by two main concepts, (1) The first one is the converged clouds and (2) The later one is the Co-Cloud.

This is why for their Clouds they are looking for FCAPS and Maintenance solutions that are scalar in a vertical agnostic fashion

To draft a business solution for FCAPS and Unified Maintenance in this scenario is very complex because traditionally each of the solution requires a tenant based analysis, processing and output which can be further customized based on RBAC functions .The task becomes more complex as we solve the myth and find that Industry leading cloud vendors have given less and not enough importance to this complex task and demarcate FCAPS boundary as the Open Stack Boundary itself which is not reasonable after all the industry is evolving from PNF to VNF and lot of points need to integrate well to replicate same or better maintenance functions of IT and Telco converged Network

There are two schools of thoughts prevailing in industry till date as per the OpenStack latest standing as per Newton Release and SDN Maruno release .

School 1: Uses the Silows and offer separate solution on Physical and Cloud/Virtualization layer (Huawei esight and Fusion Manager together make it place in this slot)

School 2: Uses the Cloud OpenStack services to unify the relation and offer complete aggregated FCAPS from single Pane of services of Telemetry like Ceilometers .The idea is good but practically till date it is a naïve idea gained less attention in the industry. In fact till now the clear mapping between two layers no IT vendor can clearly explain. In fact within the community there have been discussion on this point with no clear solution to help IT identify and solve physical alarms easily using Cloud services (Huawei System Integration Services is key to customize solution in this part)

Customer in converged environments require real time Push solutions based on API’s for key metric like Current Status of DC ,Resource usage , Resource Utilization , Tenant based metrics , Log analysis

NFV industry Defined as “monitoring as a service,” Monasca in a large framework that incorporates all aspects of monitoring, including alarms, statistics, measurements and more, for all OpenStack components. Monasca largely focuses on allowing tenants to define what should be measured, what statistics should be collected, what should trigger an alarm, and the best notification method. The relationship between Monasca and Ceilometers is currently being defined.

However still Monasca is a very vague idea , since last MWC Barcelona the Community is making many efforts in the Doctor Project and most of this effort try to replicate the existing deployed tools in the IT to the Converged Datacenters like Zabbix , Nagios , Zenoss however the fact is that they can not offer the complete solutions , this is why the industry now have a consensus to unify the Tools upstream like all of above three can be aggregated in a Open Source Project like Sensu , Similarly for Perforamcne Collectd and for logs Fluentd and the list will go on like this . https://wiki.openstack.org/wiki/Operations/Tools

The point is that in the current wave of NFV and SDN Projects the CIO sees to reduce so many tools to list of 3-5 to control the support, bug fixes and other issues that seem to be really big issues in a commercial company. I think community is doing best but still without the summation of these tools it seems difficult.

The Next wave as I see is the Unification , last month I have met couple of senior IT executives and they all have a believe with Organizations following a digital transformation path , they think there is a way to collaborate and unify many of the information in a Single Pane form operations perspective . It will open a Funnel pipe for accurate management of the cloud in terms of correlation and root cause of operational nature and obviously can build the Big IT intelligence mechanisms for a future programmable way to do things.

We need to a consensus a seamless integration and unification is a key for programmable network otherwise micro pieces even programmed in an excellent fashion cannot deliver the promise of NFV and SDN.

Just to give an idea below picture is from https://www.sdxcentral.com/ blog on SDN main components with the inclusion of more and more open source components it still looks very curious how the issues will be consolidated and solved

???????????????????

At Huawei I believe we are almost there close to community, industry, customers and market alike to build Tribalizer for Wave1 and now we need to be more open and cooperative to form alliances with other vendors and customers alike to find the Open Solutions for industry issues of how to actually achieve the Wave2?

https://www.linkedin.com/pulse/api/edit/embed?embed=%257B%2522type%2522%253A%2522link%2522%252C%2522title%2522%253A%257B%2522localized%2522%253A%257B%2522en_US%2522%253A%2522The%2520components%2520of%2520SDN%252C%2520NFV%252C%2520MANO%2520%2526%2520OSS%2522%257D%257D%252C%2522description%2522%253A%257B%2522localized%2522%253A%257B%2522en_US%2522%253A%2522When%2520you%2520curate%2520a%2520website%2520%252F%2520blog%2520called%2520PassionateAboutOSS%252C%2520it%2527s%2520pretty%2520obvious%2520that%2520you%2520would%2520see%2520the%2520world%2520through%2520OSS-tinted%2520glasses.%2520The%2520tech-world%2520at%2520least.%2520And%2520so%2520it%2520has%2520been%2520for%2520me%2520when%2520evaluating%2520new%2520comms%2520network%2520%252F%2520service%2520concepts%2520like%2520SDN%2520%252F%2520NFV.%2522%257D%257D%252C%2522author%2522%253A%257B%2522name%2522%253A%2522Ryan%2522%257D%252C%2522provider%2522%253A%257B%2522name%2522%253A%2522Passionate%2520About%2520OSS%2520and%2520BSS%2522%252C%2522display%2522%253A%2522Passionate%2520About%2520OSS%2520and%2520BSS%2522%252C%2522url%2522%253A%2522http%253A%252F%252Fpassionateaboutoss.com%2522%257D%252C%2522request%2522%253A%257B%2522originalUrl%2522%253A%2522http%253A%252F%252Fpassionateaboutoss.com%252Fthe-components-of-sdn-nfv-mano-oss%252F%2522%252C%2522finalUrl%2522%253A%2522http%253A%252F%252Fpassionateaboutoss.com%252Fthe-components-of-sdn-nfv-mano-oss%252F%2522%257D%252C%2522images%2522%253A%255B%257B%2522url%2522%253A%2522http%253A%252F%252Fpassionateaboutoss.com%252Fwp-content%252Fuploads%252F2014%252F06%252FSDN_NFV_Drawing_v2.png%2522%252C%2522width%2522%253A1096%252C%2522height%2522%253A632%257D%252C%257B%2522url%2522%253A%2522http%253A%252F%252Fpassionateaboutoss.com%252Fwp-content%252Fuploads%252F2017%252F01%252FECOMP.png%2522%252C%2522width%2522%253A602%252C%2522height%2522%253A538%257D%252C%257B%2522url%2522%253A%2522http%253A%252F%252Fpassionateaboutoss.com%252Fwp-content%252Fuploads%252F2014%252F05%252FMasteringYourOSS_CoverImage.jpg%2522%252C%2522width%2522%253A400%252C%2522height%2522%253A500%257D%252C%257B%2522url%2522%253A%2522http%253A%252F%252Fpassionateaboutoss.com%252Fwp-content%252Fuploads%252F2015%252F06%252FOSS_LandscapeTechSolnOrg_June2015.jpg%2522%252C%2522width%2522%253A170%252C%2522height%2522%253A220%257D%252C%257B%2522url%2522%253A%2522http%253A%252F%252F1.gravatar.com%252Favatar%252Fa564e905c3c15fc7268bd86149e6626b%253Fs%253D68%2526d%253Dmm%2526r%253Dpg%2522%252C%2522width%2522%253A68%252C%2522height%2522%253A68%257D%255D%252C%2522data%2522%253A%257B%2522com.linkedin.treasury.Link%2522%253A%257B%2522url%2522%253A%2522http%253A%252F%252Fpassionateaboutoss.com%252Fthe-components-of-sdn-nfv-mano-oss%252F%2522%252C%2522html%2522%253A%2522When%2520you%2520curate%2520a%2520website%2520%252F%2520blog%2520called%2520PassionateAboutOSS%252C%2520it%2527s%2520pretty%2520obvious%2520that%2520you%2520would%2520see%2520the%2520world%2520through%2520OSS-tinted%2520glasses.%2520The%2520tech-world%2520at%2520least.%2520And%2520so%2520it%2520has%2520been%2520for%2520me%2520when%2520evaluating%2520new%2520comms%2520network%2520%252F%2520service%2520concepts%2520like%2520SDN%2520%252F%2520NFV.%2522%252C%2522width%2522%253A-1%252C%2522height%2522%253A-1%257D%257D%257D&signature=AabES296V41K11DCMgJo9iqC1zcB

The Technical task is daunting but the direction is clear. Openness, alliances and sharing is key to success and together with our partners we want to build the Future Network of customers

References:

https://wiki.openstack.org/wiki/Operations/Tools

https://sensuapp.org/docs/0.26/overview/architecture.html

http://collectd.org/documentation/manpages/collectd.conf.5.shtml#plugin_aggregation

http://www.fluentd.org/

https://github.com/centos-opstools/opstools-ansible

Advertisement

Reshaping the Telco Industry of future through Service Innovation and Simplification

Sheikh is Huawei Middle East Senior Solution and Delivery Architect covering domains of NFV, SDN ,Cloud,Internet of Things (IoT) and DevOps based innovation ,specially interested in those disruptive technology driving the industry transformation

Reshaping the Telco Industry of future through Service Innovation and Simplification

“Change your head (Mind) or wait for its removal”

According to the World’s Economic Forum report 2016-17 the investment in the global Telecom industry has underpinned an immense shift in the information, cash flows and societal development mostly adding value the global economy and improving lives of masses

In fact in next 5-10years the real value that society can reap due to Digitization and information will depend on the Telecom industry provide the underlying infrastructure and improve on customer experience side but the fact is that the Telecom industry profit has declined from 58% in 2010 to 44% in 2016 end (As per survey) and it seems clear the Health of this industry is a question in itself.

Infact this issue in telecom industry is like a iceberg theory and many coupled issues so CEO need to solve these underlying issues instead of just barraging old themes of aggressive products and rollout to improve profits .In fact such an approach can only bleed the industry woes to the bottom line . Traditionally IT and Telecom industries considered separate but the fact is that the silent nimble players like Skype, WhatsAPP, IMO , Wechat has silently taken huge profits while using Telco infrastructures to reach customers merely through digitization of services

Below are some key actions to focus, not only ideas some on ground fact that will put the Telco’s back in command

Viz.

1.    Re-Architecting the Future Networks

Network synergy, Energy efficiency, Infrastructure sharing, high capacity OTN networks, Single RAN, Virtualization, NFV, Cloud based. These are all the terms you will listen from time to time but theme is simple make architecture simple and modular

It will enable you to not only offer a competitive price but also make new services innovate easily

2.    Operations simplification

This is a very important target for industry. I remember the ETSI Phase2 specification specially considered the integration between NFV and SDN to solve enterprise needs for use cases that extend beyond geographies of a typical data center.

Similarly the internet style of operations where main skill is based on programming and Meta modeling of issues instead of troubleshooting skills by written guides is very important.

In Huawei we have developed a customized process for the Telco’s based on DevOps but meeting more specific needs from carriers like high quality testing , fault demarcation and continuous delivery making best product based on Pilot and Production environment synergy and concept of remote delivery . We call this process IntOps.

The key point is that Telco’s can not merely copy IT companies default DevOps, Chef and Ansible playbooks to use them directly.

I think this is the main theme of CTIO to think how to achieve this goal.

3.    Orchestration

What is our definition of this metric is how the experience will be delivered to customer through Omni channel. The standardization of Os-Ma interface will lead us to a world where customer can not only purchase but can drive the service creation through online experience.

The Service Logic and Open Graphs will be intelligent enough to create services across many infrastructures

4.    IoT

The internet of everything is a game changer, and due to the reason each IoT use case is customized in terms of QOS and service modeling makes Operators prime drivers to speed up and monetize this opportunity for their transformation.

5.    Security ,Cyber crimes and personal data integrity 

As data volume increases beyond imagination and networks become converged the threats for safe communication is immensely threaten by the Security guarantees and the intrusion detection.

·        Security policy interaction with security devices

·        Technology to defense APT attack

·        Utilizing Industry Best Practice (CIS Standard) to enhance OS

·        Security abilities more than the original OpenStack

·        Security isolation of Virtual Resource for VNF

·        The integrity protection based on TPM

·        Introspection technology for business information encryption

·        Secured deployment, upgrading and PRIST

·        Actively participate in ETSI NFV Security Specifications

6.    Organization Re-structuring

Who will solve this issue when the decision maker itself fears his redundancy? Infact this point is complex with Telecom aging workforce. It seems difficult to tell a person all the skills he acquired in two decades are not relevant now.

How to let people work in cross function Scrum teams, how to replace each manger by leader or problem solver.

To solve idea is simple, nothing can be achieved overnight. Let us promote guys who accept the change and upgrade. Set targets and make policies that discourage lazy team’s .Focus on software skills more. Python /Cheff /OpenStack /JSON

7.    Application or Infrastructure make clear decisions

Compared to IT industries, still the Telecom industry reliance on infrastructures is huge .By this I means the Telco services are not Infrastructure independent. While on Enterprise Architecture for IT the whole pie lies on the Application itself.

NFV micro service is all the stuff targeting Telco separate from infrastructure but fact is that still many services rely on his part like SR-IOV, Smart NICS, FPGA and accelerators

The strategy set on this level will have huge impact on the future of Operators. The industry leaders like AT&T , Verizon , NTT ,DT all have make clear plans to be the Software defined operator so I think actions should follow the same .

8.    Replace troubleshooting by service automation

This is a crazy idea. The SON and Software defined networks can scale and heal automatically. But the programming logic and map need to build by operations team.

It require more specialized skills on both service modeling and coding and obviously many operations jobs will be depleted .

9.    Co-deployments and converged clouds

Telco in recent years have already deployed Cloud based applications like Cloud Core, Cloud Edge, VAS Cloud but the fact is that customer CTIO must view Cloud as a shared pool which is independent of any application. The year must see ways how to join the clouds.

Till now the Global resource sharing among tenants is key point that is not standardized and difficult to customize based on each tenant. I think the community is not doing business driven approach to solve such issues. This is also a reason why commercial companies look for productions ready Distros like Huawei Fusion, Mirantis Ocala MOS, and Redhat OSP etc

10.                       Process mapping 

This is last and most important part, as from decades Telco’s have followed a very standardized process and structures to run the organization. with the digitization many of these things is going to change like Team structures , process , procedures , upgrade the eTOM processes to the ITIL based processes , Define SLA’s across IT and Telecom domains and incorporate orchestration and key enabler

I have this idea to make the processes the Scrum leader must be appointed and below team member is essential to map the processes. It is good idea to match the processes with PMI ® ACP XP (Extreme programming matrix)

Based on across industry analysis it seems evident that with huge challenges ahead the time is now to adapt .On operational level its more about transforming the ways of working and on strategic level is to deliver complete digital experience ln addition we need to solve the cross industry Silows and specially legal and regulatory concerns

According to the World Economic Forum report there are still other factors like micro branding, atomization, legal and lastly digitization of user experience side but since the scope is too wide and I am not expert on such topics, so I am leaving those for other industry experts.

I have a sincere hope that these 10 points are very important for CXO to drive the Digitization of services across industries. Future we will not listen Cloud , NFV , SDN as different terms , also industries like banks , Telco’s , enterprise not separate Isle’s but joint realities driven by the common goal of digitization

Reference:

1>ETSI standards

2>ITIL expert boot camp

3>World Economic forum report Released in January 2017

https://www.linkedin.com/pulse/reshaping-telco-industry-future-through-service-saad-sheikh

 

 

 

 

How to plan the Best Server Solution for a Converged Telco Cloud Data center

During a recent workshop one question popping up which server make is the best to be deployed in a Telco Data center ?

As a Solution Advisory i know that such a question has no fixed answer and answer depend on solution , its basis and long term business and network strategy of customers but still for such questions customers require explain in terms of experience in similar projects and how to actually make such a comparison ? Specially for converted IT and Telco Data Centers .

To start with for this scenario in carriers business we have two solutions viz. Rack Servers and Blade Servers to fill in the blank ; the first thing we need to look at is the industry of NFV and Enterprise till date .What we know is that in NFV normally we are using 42U racks and Blade servers like HP7000, C7000 Gen8/Gen9 Servers, and Huawei E9000 all are blade servers and also Dell M630 also a blade server.

Technically we all know both server types are COTS X86 machines so must be OK to use any of them then why industry has come to such thinking of adopting the Blade servers? After detailed research and compared the benchmarks I can summarize the selection based on following items

1.Performance: in a Telco Cloud, performance is the killer decision because customer will only adopt the solution if it is same or better than the legacy one. We all know that blade servers processing in terms of V3 CPU Core, Disk and Obviously DRAM is much faster than Rack. Although the advocates of Rack servers proclaim with new technology the expansion in upgrade in rack servers is easy but it is balanced when we know the Space /Power /Maintenance of cabling issues in Rack server deployments

2. Cost: There is no doubt the Rack servers main intention was to support mass deployment in data centers which require less performance and quality, For example Google Cloud, Amazon Web hosting, Face book .In fact for these data centers it is more pragmatic to make expansion in bay and for complete racks to have cost advantage, this is one area where Rack servers are certainly doing better than blade servers

3.Scaling: a Telco Cloud must be hyper scalar .It means we can increase by server, by module without complexity. In fact for both type of servers the Compute node expansion is possible by scalar way but the complexity for O&M and installation for the two is not the same. We all know Rack servers need more cabling, more artifacts to get the job done .So for me it lacks promise of scaling

4.Energy: In a big data center it matters most to save power. Comparing the different solutions from different vendors we found that the Blades are more compact. I knew there is a term famous to data center guys about Power usage like power provided to application specific nodes to total power input. Sorry I fail to recall the ratio name but it is far better in Blades in Telco Special case

5.Infrastructure cost: The blades are best from Form factor point of view with Shared power, switches, disk and is a reason for less cabling needs compared to Rack

6 Solution unification: I think all Solution Architects know in a complex ICT world all the big carriers are trying to consolidate both IT and Telco domains to come to a unified solutions Finally .the initial High CAPEX is OK if it promise large OPEX reductions in the TCO cycle. Since Blade servers promise certain optional features and functions which Racks do not have so it is reasonable to deploy with Blade servers or to select multiple type of rack server’s .To explain this one can see the open Stack EPA (Enhance platform awareness) case what we know is that the ideas of Solution unification fits well for NFV when we use Blade servers

I hope these bench marks will enable you to plan your data centers but please note I am enthusiastic writer and blogger. So the comments and analysis in this paper are Sheikh’s own as per industry reach and self understanding, they do not represent any official understanding from my Employer Huawei.

Secondly there are many white box vendors and market of both Rack and Blades are huge. You may come to a situation where the Blade in analysis is worse than the counterpart so the reviews are more relevant if you are planning a Telco Cloud because this is my main focus area…

Adoption challenges in Open source and opportunities for niche System Integrator

In 2019 the IT and Telecom industry has really come to synergy point through the power of open Source solutions. I still remember days in September 2012 when we are firstly looking the Open Source solutions .It was really difficult that time to even think the Open Source can integrate the so called real Isle Industries of IT and Telecom. However the fact is that big Companies already doing production ready work for consolidation of IT and Telecom Datacenters.

What is most exciting for me is that I started 2019 with a CXO workshop asking the way forward to include the Banking Application on the same Data Center. Obviously at this time the idea is very vague because the Banking as we all know requires highly secure and reliable. Many of the ideas of traditional IT and Telecom cannot be directly applied there .However the message is clear the future IT services will cover whole Ecosystem and infact each industry is Just like Application appliance .

The Cloud Based infrastructure is an idea which is so important to develop in multi tenant environment because each vertical will come with its own set of requirements. It is clear that there is huge market demand for this and actually no vendor can eat all this market though many will try to do this.

This type of Cross industry and Multi tenant solutions will strongly depend on the power of Open Source solutions to couple the solutions form many vendors as a module. The concept which is known in industry as a micro service .However the integration of all these Open Source is very complex especially when it comes to adaption and matching CIO requirements. It is because although the solution can be Open but customer requirements can be customized. It is because no customer want to live in a house build on the wish of someone .The Personalization of each customer solution is very important.

Obviously all customer wish to transform to an agile infrastructure and reduce OPEX but really what must be there first steps like how to select the Components from the Open Source Solution? What must be there Cloud, VIM, MANO, VNF, Repository etc. This type of question many CXO already put to vendors and the answer of each vendor depend on its business model like what is there Ecosystem without delving to details whether it can satisfy customer real needs

I think this strategy is damaging the NFV adoption on a wider scale, because vendors want to push a solution while the best strategy for open Source solution must be a Pull solution which matches customer needs. Without this it is difficult to increase the real demand of NFV and Open Source Projects

After selecting the software the next daunting task must be how to make it functional.  The product view of this scenario is to develop all applications as a VNF Appliance through image file. It means any application can be available and only usage is licenses. However the story is not so simple. Infact it is in stark contrast to the Idea of App Store as prevailing in IT industry.

The part of such endeavors must be how to efficiently use these images to develop a service, something customer really need, way to auto test and auto deploy the service and obviously how to apply Puppet, Ansible and Chef Play books schemes to achieve it. I mean to say just the Appliance and their abstraction is not enough how to model the service and make it interwork is key to success for this realization

This type of open Solution ecosystem will support the wide range of adoption across the industry; maybe we will see the NFV going live in both Private and Public Clouds and the application will be as freely distributed as the Windows of today

However with more and more open solutions the dilemma to integrate them will become more and more important question by all CIO’s , not only this we will come to a situation where customer will question the reliability and performance of open source solutions in commercial deployments ? Can SI can use all open Source tools and offer something a single Pane glass operations modeled around Open API’s and programming abstraction .

This is the real value of System integrator, the Hercules the IT and Telco industry waiting for years.

Can 2019 be the year of that SI’s making an impact? The story is open and the best team will make it happen. System integration team has to listen to customers listening to the industry, partners and customer alike to solve the issues of Ecosystem.

Collaboration is the key and sharing of experiences to solve industry issues is critical to be the System integrator of choice for the industry

https://www.linkedin.com/pulse/nascent-open-source-solutions-success-looking-si-saad-sheikh-be-pe