José Ribamar Smolka Ramos
Telecomunicações
Artigos e Mensagens


WirelessBrasil

Março 2012               Índice Geral


03/03/12

• Mensagem de José Smolka (texto em inglês): "NGN and traditional telecom vendors"

de J. R. Smolka smolka@terra.com.br por yahoogrupos.com.br
para "Celld-group@yahoogrupos.com.br" <Celld-group@yahoogrupos.com.br>,
"wirelessbr@yahoogrupos.com.br" <wirelessbr@yahoogrupos.com.br>
data 3 de março de 2012 20:21
assunto [wireless.br] NGN and traditional telecom vendors

Hi people,

First of all let me explain wy I'm writing this in English. It's just because I feel more comfortable with this language when I'm really pissed off. If I try to write this in Portuguese it would probably have to be blocked by our dear group moderators because of my liberal use of inadequate words.

Next question that raises is: why am I so pissed off anyway? That happens every time I see how traditional telecom equipment/solution vendors behave when selling IP-based networks, which we conventioned to call Next-Generation Networks (NGNs).

Yesterday there was a message sent to celld-group, posted by Cesar Nunes, making a few questions about NGNs. I've answered this message, but that's beside the point here. What really interested me was another reply, sent by Juliano Correa, which got an Ericsson presentation attached. Ericsson is *the* traditional telecom vendor, and some points in that presentation slides did get me wanting to kick Ericsson's ass. And so I'm doing.

Those slides were made - so it's said on the first slide -  for a presentation at the Advanced Level Telecom Training Centre - ALTTC, at the city of Ghaziabad, Uttar Pradesh province, India, about the network changeover being planned (and done?) at Bharat Sanchar Nigam Ltd. - BSNL, an Indian carrier headquartered at New Delhi, back in 2006.

Don't know how much of the network transformation outlined in those slides were really put to carry real user's traffic, but there's something to say for the network planning & design people from BSNL: not many carriers were seriously contemplating so huge a network transformation at the time. Comes to mind British Telecom Plc and it's BT21CN project, and that's all.

Very well. So what's about those slides to anger someone like me? There's two points, very characteristic of traditional telecom vendors of the time - and today, which I feel truly irksome.

The first one is on the second slide, where there's the following phrase: "Carrier class Telephony over IP – just VoIP not good enough". There are few expressions that can put my patience in overdrive, but the use of "carrier class" as an adjective to packet-switching solutions surely is one of them. This implies that carriers' networks hover in another level of seriousness and quality when compared with native IP solutions. Sorry, but this is one of the cases where I see no option, except being rude. I have to say: carrier class my ass.

I know this situation quite well. Back in the 1980's we (me included), IBM mainframe professionals, used to say quite the same derogatory things about using personal computers and LANs to build real-life, heavy-duty, mission-critical computing systems. They just looked as interesting toys, but using them to do real work? No way. Well, what really happened is history now. Even me, seasoned IBM mainframe systems engineer (I used to work with VM/XA and MVS/XA at the time), had to recognize the changing, volunteer to be at the cutting edge of the changes, and set that our nascent TCP/IP corporate network service level should be no less than that provided by the IBM SNA network (wich was mainstream then). Either that or not receiving users' trust to change their applications to the new network.

I remember that, when the project was still in design phase, we visited some enterprises known for running big (by the standards of that time) TCP/IP corporate networks. I told of this goal to one of our hosts, which already had about 10 years of experience in LAN networking and TCP/IP internetworking, and she laughed at me, saying that it was impossible to achive at the time. But we did it. One and a half year after that conversation our corporate TCP/IP network had about 1200 connected hosts, including the IBM mainframe, and users could use their personal computers to access legacy applications through TN3270 and also use personal productivity applications (spreadsheets, text processors and so on).

Traditional telecom vendors and carriers' network engineers posture about VoIP is very similar. And will end the same way. Carrier's engineers have an excuse at least: carriers' O&M work environment isn't very inviting for innovative thinking. One could think that carriers'design & planning people would do better, but - surprise! - they don't. I've been asking around and found that almost all carriers don't have a network modernization rodmap of their own. They simply take the easy way out: rely on their vendors' roadmap. I know it. Been there, seen it from inside, found it un-fu**ing-believable, talked to everyone I could about it. But made no difference. And that's why I feel that posture from traditional vendors so unbearable. For their own business reasons (they still have a cash-cow to milk away, after all) they've been instilling FUD over carriers' people - and still do so today, and stalled network evolution for at least a decade.

And the second irksome thing I've found is at slide 15. There's the information that IP routing among the various elements of the proposed NGN solution (softswitches, IMS, etc.) would be borne by Cisco routers. That's nothing remarkable per se, but they add the following commentary: "carrier-class implementation meeting the telephony requirements". I've seen some of those plain vanilla Cisco routers & switches implementations with the "carrier-class" attached to them. Why so?

Looks like the idea is to convey the feeling that this kind of equipment is better engineered than those which Cisco would sell to the rest of their clients. Well... unless you think that buying the routers and switches with redundant DC power sources makes them "carrier-class implemented", that's an outrageous lie. And ever have you guessed why the "carrier-class" routing & switching solutions are always top of line equipments? Cisco's 12K and CRS-1 routers still are top of line today. Just imagine how unattainabe they looked back in 2006. But i suspect there's a reason behind this.

The benevolent interpretation is that they simply don't know how to make capacity planning for this kind of box, so they prefer to err on the plus side. But there's a malevolent interpretation too: specify the top of line boxes imply that less capable IP routing equipment won't be up to the task of carrying telecom traffic. That's a huge bullshit. And there's the possibility of the two interpretations being true. Ain't that beautiful?

I really hope that this message could spark some kind of debate in the discussion groups - otherwise they won't be "discussion" groups anymore. It's kinda lonesome keep talking without any feedback y'know? Wake up people! Say what you think! What the fu** are you doing out there? Sleeping? I may be acid and rough sometimes, but I don't bite. And even if you don't agree with me, it's your right to do so, and to tell it to anybody who cares to hear.

Let's wake up people!

[ ]'s
J. R. Smolka


WirelessBrasil                     BLOCO