Lessons from NFV de­ploy­ments

Ev­ery role has dif­fer­ent re­quire­ments and pri­or­i­ties, there­fore the touch points, chal­lenges and lessons of NFV im­ple­men­ta­tions can be starkly dif­fer­ent

Voice&Data - - CONTENT - (The author is CEO, Open­wave Mo­bil­ity) john giere

1. CXO

The big change for the CXO is the shift from CAPEX to OPEX. Cloud-based, on­de­mand, vir­tu­al­ized in­fra­struc­ture is an OPEX play, while the tra­di­tional ser­vice provider play has been on cap­i­tal-in­ten­sive, hardware-ori­ented ex­pen­di­ture.

In our ex­pe­ri­ence, this shift has not been fully rec­on­ciled within many ser­vice providers. While it may seem ob­vi­ous to busi­nesses that op­er­ate SaaS busi­ness mod­els or the many busi­nesses that have come to rely on SaaS, this is quite a ma­jor shift in think­ing and pro­cesses for ser­vice providers. Do not un­der­es­ti­mate how chal­leng­ing it is to change “the way we’ve al­ways done it”. It can be just as com­pli­cated as any tech­ni­cal chal­lenge.

As time goes by, we ex­pect 30% of the OPEX to go to NFV in­fra­struc­ture and 70% to go to VNFs/VNFCs. This is the first les­son - that this shift to­wards OPEX needs to get un­der­way or it will create an in­sti­tu­tional bar­rier to suc­cess­ful NFV de­ploy­ments.

2. CFO

From the CFO’s per­spec­tive, the les­son is on ROI from an NFV in­fra­struc­ture. The tra­di­tional mind-set has been on ROI from phys­i­cal in­fra­struc­ture. A vast sim­pli­fi­ca­tion is that if we add X phys­i­cal in­fra­struc­ture, we can de­liver Y rev­enue. But this think­ing is not suf­fi­ciently lay­ered for an NFV world.

The ROI will only be break-even if it is based on phys­i­cal in­fra­struc­ture. With NFV, it is about be­ing able to do things bet­ter, rather than hav­ing bet­ter things. The true pay­back is based on the change of process as well as in­fra­struc­ture.

3. En­gi­neer­ing

En­gi­neer­ing has quickly caught up to the need for new skillsets - there is no new les­son there. The les­son here is that you

ac­tu­ally don’t need to be at the bleed­ing edge, as many may think.

This is fairly typ­i­cal or­ga­ni­za­tional think­ing. When peo­ple within an or­ga­ni­za­tion re­al­ize that they are be­hind on some­thing, the hu­man ego of­ten kicks in with some­thing like “We’re ter­ri­ble at this, but we should be the worlds most ad­vanced at this!” And this think­ing can fil­ter down to KPIs and dic­tats. New skills ARE needed for NFVI, but you don’t have to be on the bleed­ing edge. Out there in the real-world, we find that most OpenS­tack op­er­a­tions are based on the older Mi­taka, rather than the newer Pike. And that’s okay.

4. Prod­uct Man­age­ment

For prod­uct man­age­ment, the les­son is about mak­ing a choice with their eyes open. With the im­ple­men­ta­tions we’ve been in­volved in, it’s been 60% OpenS­tack and 40% VMware.

The de­ci­sion here is much more ex­pan­sive than can be fit into a bul­let point. But at its core, it’s an op­er­a­tional de­ci­sion and a cost de­ci­sion. To my mind, while be­ing care­ful not to wade into the re­li­gious fer­vour of smart­phone OS pref­er­ences... it has a par­al­lel in the choice be­tween iOS or An­droid, more so a few years ago.

Do we go for a more costly but ma­ture sys­tem with ad­vanced tools or a flex­i­ble but ma­tur­ing open en­vi­ron­ment? That is a big de­ci­sion for any En­gi­neer­ing team - and one that needs to be made un­der­stand­ing the full, long-term im­pli­ca­tions of both paths.

5. Ops

As for Ops, their les­son is both chal­leng- ing and op­ti­mistic. Vir­tu­al­iza­tion can and should have a dra­matic, pos­i­tive im­pact on pro­vi­sion­ing speed. We con­sider that Ops need to be look­ing at a 50%+ sav­ings on the process time for ser­vice setup and pro­vi­sion­ing.

This bench­mark will en­sure that the re­al­i­ties of process im­prove­ment will start to pos­i­tively im­pact the ROI from vir­tu­al­iza­tion. Plus, it is not so am­bi­tious that it will have a neg­a­tive im­pact on the Ops team.

6. CTO

To mis­quote “Top Gun”, the CTO is go­ing to feel the need, the need for speed. Ser­vices and cloud con­nec­tiv­ity have to be low la­tency. It will not be ac­cept­able for there to be an in­crease in la­tency within an NFV en­vi­ron­ment.

In the tele­com world where five nines avail­abil­ity has been a mantra for decades, it cer­tainly won’t be ac­cepted in an in­fra­struc­ture that is meant to be an up­grade. The tar­get for 5G en­vi­ron­ments is line speed. As an ex­am­ple, for Open­wave Mo­bil­ity our vGiLAN func­tions work at sub-5mS and we are con­stantly work­ing to im­prove that.

7. Cus­tomer Ex­pe­ri­ence

And let’s not for­get who tel­cos are do­ing this for. The peo­ple, that we all ul­ti­mately serve - the peo­ple who pay the bills, the subscribers them­selves.

For the peo­ple whose role it is to en­sure and im­prove the cus­tomer ex­pe­ri­ence, their ex­pe­ri­ence of NFV will be dif­fer­ent again. For peo­ple in th­ese roles, our ad­vice would be to con­cen­trate on three clear met­rics to mea­sure. It will likely be sub­tly dif­fer­ent for ev­ery op­er­a­tor, but for ex­am­ple it could be im­proved through­put, re­duced RAN con­ges­tion or im­proved Qual­ity of Ex­pe­ri­ence.

What­ever the met­rics are, they need to be mea­sur­able and man­age­able. And from their per­spec­tive, an NFVI has to serve them, so they can bet­ter serve their subscribers.

Mass-mar­ket NFV is be­com­ing more of a re­al­ity day by day. The fu­ture is bright, even if it’s more com­pli­cated than we would like. Thank­fully, from the im­ple­men­ta­tions so far, there are al­ready lessons that can be taken, what­ever your job role.

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.