I’ve just returned from a European trip, one that gave me a chance to dig a bit more into service provider visions and even economic views. I think that there are changes, perhaps even massive ones, on the horizon in both areas.
On the service provider side, I’ve found a dramatic shift in the way that senior management views technology tools. The view of “the network” as a profit center is quickly eroding in Europe, where competition, regulation, and economic conditions combine to make traditional services even less profitable than they are worldwide. Five years ago, the CTO of operators was focused totally on network technology. A couple of years ago, they added a server focus. Now, they see servers and software as the crux of their tech future, with the network only the connecting tissue.
This isn’t going to bode well for equipment vendors whose futures were already in doubt given the last quarter. It’s pretty obvious that none of the major vendors are optimistic, at least not in the privacy of their planning sessions, and obvious they have no reason to be. Even current trends in revenue per bit were guaranteed to commoditize bandwidth and the stuff that creates it. Now we’re seeing, in the SDN and OpenFlow model, a specific way to make the network the slave of the cloud. It may be that SDN and OpenFlow have catalyzed the whole cloud shift, in fact.
This isn’t to say that network vendors are in immediate jeopardy. So far, our spring survey isn’t showing any massive net loss of strategic influence among the network vendors, nor any significant gains among IT giants. The most software-server-centric of the group—Cisco first and perhaps Alcatel-Lucent second—have plenty of opportunity to position their assets, but it’s not clear whether they have the desire to do that. Alcatel-Lucent, after all, announced it was getting into the core router business, which suggests that it’s still counterpunching with Cisco and Juniper (and Huawei) for bit-pushing market share. Cisco’s comments over the last couple of months suggest it might be seeing the light in differentiating the network based on how well it supports software apps, but I’ve not seen much new here. Cisco also issued its all-too-regular prediction of congestion doom, suggesting operators buy routers even if they can’t make money just to be nice to users.
Alcatel-Lucent announcement may be more a PR failure than a product strategy failure. The operators tell us that they perceive the new 7950XRS as being aimed at the interconnect points and even metro apps (see more on this below). While the press release does say that the new router is targeted at boundary points and the cloud, the “enters the core router market” tag makes it hard to see past the market hype.
Cisco’s delay in making its SDN/OpenFlow position clear may be due to the fact that it knows the announcement is important and has to get it right. In fairness, the theme of my blog today is in part the fact that operators have suddenly transformed their view of the future. Vendors can be excused for taking some time to catch up—but only “some” time. Cisco will need to sing better if it sings later.
On the topic of operator vision, operators have been saying a lot of revolutionary stuff even in public. Vodafone has said that its CTOs are likely to be more computer types than network types. Verizon says that scaling Ethernet is their main reason for increasing the use of MPLS in the metro network. I’ve been telling vendors for years that the real value of MPLS, even of IP, has to come in the metro network because that’s where profitable traffic is found. The IP core is the value sink of bits so the less you spend there the better. This is why it’s critical for operators to understand how OpenFlow fits; it may be the alternative to both Ethernet and MPLS in the metro, particularly if that elusive service/experience linkage to OpenFlow can be found or created.
The court decision that APIs cannot be patented might have significant ramifications in this area. Before the ruling, my EU friends were saying they believed that the fact that the EU was taking that position meant that it had a more software-friendly attitude and might in fact be the software innovation center for the future. With the US also in that camp (pending inevitable appeals) the question now may be whether the fact that APIs are open by definition means that we might see more cloud competition. Google won with Android APIs but might lose in the longer term because it may find it impossible to prevent others from duplicating cloud-service interfaces Google offers. Same with Apple. This is good for the market, I think, because it would prevent any giant from stifling innovation by locking up key service practices in the form of copyrighted APIs.