Is IP Growing UP? Is TCPOSIP the New Protocol Stack? Will Sessions Pay For Networks?
Oracle’s purchase of leading SBC vendor (session border controller Acme Packets), is a tiny seismic event in the technology and communications (ICT) landscape. Few notice the potential for much broader upheaval ahead.
SBCs, which have been around since 2000, facilitate traffic flow between different networks; IP to PSTN to IP and IP to IP. Historically traffic has been mostly voice, where minutes and costs count because that world has been mostly rate-based. Increasingly they are being used to manage and facilitate any type of traffic “sessions” across an array of public and private networks, be it voice, data, or video. The reasons are many-fold, including, security, quality of service, cost, and new service creation; all things TCP-IP don't account for.
Session control is layer 5 to TCP-IP’s 4 layer stack. A couple of weeks ago I pointed out that most internet wonks and bigots deride the OSI framework and feel that the 4 layer TCP-IP protocol stack won the “war”. But here is proof that as with all wars the victors typically subsume the best elements and qualities of the vanquished.
The single biggest hole in the internet and IP world view is bill and keep. Bill and keep’s origins derive from the fact that most of the overhead in data networks was fixed in the 1970s and 1980s. The component costs were relatively cheap compared with the mainframe costs that were being shared and the recurring transport/network costs were being arbitraged and shared by those protocols. All the players, or nodes, were known and users connected via their mainframes. The PC and ethernet (a private networking/transmission protocol) came along and scaled much later. So why bother with expensive and unnecessary QoS, billing, mediation and security in layers 5 and 6?
Then along came the break-up of AT&T and due to dial-1 equal access, the Baby Bells responded with flat-rate, expanded area (LATA) pricing plans to build a bigger moat around their Class 5 monopoly castles (just like AT&T had built 50 mile interconnect exclusion zones in the 1913 Kingsbury Commitment due to the threat of wireless bypass even back then, and just like the battles OTT providers like Netflix are having with incumbent broadband monopolies today) in the mid to late 1980s. The nascent commercial ISPs took advantage of these flat-rate zones, invested in channel banks, got local DIDs and the rest as they say is history. Staying connected all day on a single flat-rate back then was perceived of as "free". So the "internet" scaled from this pricing loophole (even as the ISPs received much needed shelter from vertical integration by the monopoly Bells in Computers 2/3) and further benefited from WAN competition and commoditization of transport to connect all the distributed router networks into seamless regional and national layer 1-2 low-cost footprints even before www and http/html and the browser hit in the early to mid 1990s. The marginal cost of "interconnecting" these layer 1-2 networks was infinitesimal at best and therefore bill and keep, or settlement-free peering, made a lot of sense.
But Bill and Keep (B&K) has three problems:
- It supports incumbents and precludes new entrants
- It stifles new service creation
- It precludes centralized procurement and subsidization
With Acme, Oracle can provide solutions to problems two and three; with the smartphone driving the process. Oracle has java on 3 billion phones around the globe. Now imagine a session controller client on each device that can help with application and access management and preferential routing and billing etc, along with guaranteed QoS and real-time performance metrics and auditing; regardless of what network the device is currently on. The same holds in reverse in terms of managing "session state" across multiple devices/screens across wired and wireless networks.
The alternative to B&K is what I refer to as balanced settlements. In traditional telecom parlance, instead of just being calling party pays, they can be both called and calling party pays and are far from the regulated monopoly origination/termination tariffs. Their pricing (transaction fees) will reflect marginal costs and therefore stimulate and serve marginal demand. As a result, balanced settlements provide a way for rapid, coordinated roleout of new services and infrastructure investment across all layers and boundary points. They provide the price signals that IP does not.
Balanced settlements clear supply and demand north-south between upper (application) and lower (switching,transport and access) layers, as well as east-west from one network or application or service provider to another. Major technological shifts in the network layers like open flow, software defined networks (SDN) and network function virtualization (NFV) can develop rapidly. Balanced settlements will reside in competitive exchanges evolving out of today's telecom tandem networks, confederation of service provider APIs, and the IP world's peering fabric, driven by big data analytical engines and advertising exchanges.
Perhaps most importantly, balanced settlements enable subsidization or procurement of edge access from the core. Large companies and institutions can centrally drive and pay for high definition telework, telemedicine, tele-education, etc... solutions across a variety of access networks (fixed and wireless). The telcos refer to this as guaranteed quality of service leading to "internet fast lanes." Enterprises will do this to further digitize and economize their own operations and distribution reach (HD collaboration and the internet of things), just like 800, prepaid calling cards, VPNs and the internet itself did in the 1980s-90s. I call this process marrying the communications event to the commercial/economic transaction and it results in more revenue per line or subscriber than today's edge subscription model. As well, as more companies and institutions increasingly rely on the networks, they will demand backups, insurance and redundancy ensuring that there will be continous investment in multiple layer 1 access networks.
Along with open or shared access in layer 1 (something we should have agreed to in principal back in 1913 and again in 1934 as governments provide service providers a public right of way or frequency), balanced settlements can also be an answer to inefficient universal service subsidies. Three trends will drive this. Efficient loading of networks and demand for ubiquitous high-definition services by mobile users will require inexpensive uniform access everywhere with concurrent investment in high-capacity fiber and wireless end-points. Urban demand will naturally pay for rural demand in the process due to societal mobilty and finally the high volume low marginal cost user (enterprise or institution) will amortize and pay for the low-volume high marginal cost user to be part of their "economic ecosystem" thereby reducing the digital divide.
TechZone 360 Analyzes the Deal
Acme Enables Skype Bandwidth On Demand