- The year of the paradigm change -

Our best wishes for you and your beloved ones!  
In this new year of the Dragon, we see the promise of a critical change in the approach towards IP data growth and consequences. And the most noticeable change will probably be the paradigm change from volume to real time bandwidth to manage Quality of Experience (QoE) and congestion impacts.

From a carrier standpoint, volume data explosion has been considered as a real threat: many tools and solutions have been deployed to support this growth and reduce congestion impacts for all subscribers. Carriers have blown the whistle on heavy users who have consumed too many GB of data. Heavy users have been clearly contributing to unfairness while consuming too many network resources. However, volume, while an indirect cause of congestion and QoE degradation, is not the most relevant criteria to manage both communication towards subscribers and congestion mitigation at the network level.

Volume is used today to communicate limits or service package thresholds to subscribers. Problem: this is a very tough exercise. Communication on a volume basis towards subscribers, to explain them what they are entitled to, should mean that users know exactly how many MB will be consumed for their HD movie on YouTube, or for their Application update on their smartphone. This requires real time measures (metering) and visibility for the subscribers.

Volume as a criterion to mitigate congestion is not relevant as network resources are getting scarcer: volume per day, or volume per month are sometimes considered as good indication about bandwidth usage. But this is in fact a static view about what is going on in the network (see for instance:  heavy users don't necessarily cause congestion claims report). 2GB of data transiting at 2am in the morning has obviously not the same impact and the same cost for the carrier, as 2GB of data transiting at 7pm in the evening during the so-called peak times. Congestion in the network is dynamic with parameters that are changing every second. And every subscriber does experience the alteration of their QoE as they are using their smartphone in real time. 

Fortunately, this static perception about volume is now changing to a dynamic approach.  The new paradigm resides in the capabilities to expose real time usage information (with online charging, or real time BI) and to measure and steer effectively real time bandwidth for each concurrent user. This definitely requires a radical technology change in analyzing and controlling network traffic flows with increased processing power to manage millions of  contexts/counters and parameters. This is what Vedicis software Content Smart Switch offers today by fully exploiting and unleashing Intel server performance.

- From Network to IT and vice versa: Policy Control & Charging and the pricing case -

 It is common in the telecom industry to talk about the IT & Network convergence, while sometime it would be more precise to talk about collision.

Interestingly enough, DPI-Policy enforcement platforms are precisely at the cross road of IT and network influences. On one side, there is a vertical integration with the control plane, with radius, policy servers and charging/billing. IT has a clear influence, and the industry has seen in the last months a couple of acquisitions and mergers in the SPIT (Service Providers IT) ecosystem that highlight the growing influence of BSS actors in this area.

On the other side, or should we say, on the orthogonal side, the user plane with routers, switches and other equipment is traditionally managed by Network teams. As far as DPI and Policy Enforcement are concerned, Network Equipment Providers for Mobile carriers and 3GPP future releases show a target architecture with DPI-PCEF features in the GGSN or PSDN GW. This is in “the roadmap”.

Today, “crossing the chasm” means the convergence of IT and network teams in order to have a holistic approach towards users and tap the full potential of Internet usage. This implies a common terminology. Did you notice that Network staff usually talks about bandwidth and capacity? Concurrently IT talks about users…

When Vedicis announces a full Policy Control & Charging Solution with FTS, the purpose is not only to enforce smoothly profiles and policies defined in a PCRF in the network. It is not simply about technology. The purpose is to bring an integrated solution that creates a common view about users from service packages to user bandwidth allocation and charging. An integrated solution means a single pricing based upon users, with their average traffic and usage.  Finally for IT and Network teams, this single user pricing gives a common basis to discuss investments and ROI with a customer centric approach.

- Software DPI & Policy Enforcement -

Lastly at the Broadband Traffic Management Congress, we were asked on various occasions to explain why Vedicis was so different regarding DPI approach.

Is it a question of performance? Flexibility? Features?

Basically, yes to each question, but above all, it is a question of nature of the technology itself that is changing completely the approach towards DPI evolution. While current DPI architectures are based upon traditional network hardware either with Network Processor Unit (NPU) or ASIC, Vedicis has designed a full software DPI technology that is completely executed on standard Intel servers. It means these server gears that are shipped by millions every year by HP IBM and others, and without specific networking cards…But why is possible now and not before?

5 years ago, Intel started to invest in its new CPU generation with a completely new architecture called Quick Path Interconnect (QPI) that was aiming at achieving new level of performances without increasing CPU frequencies/power consumption. The new CPU featured 4 cores, with 2 threads: that was the equivalent of 8 cores with large L2 and L3 caches.  At the same time, the QPI architecture proposed a direct connection to the RAM through 3 channels. This was a critical change that suppressed the bottle neck from the Northbridge linking the CPU to the RAM. And all of a sudden, the bandwidth from CPU to RAM was multiplied dramatically.

Why does it matter?

With this Intel architecture, Vedicis has developed software orchestrating correctly these CPU and RAM to process millions of flows and user connections. CPU gives high processing power, and RAM (almost unlimited) enables the management of millions of information (users, content , applications…) all the more as there is no more bottleneck between CPU and RAM.  Vedicis has built Intellectual Property and its development to take the most of this new architecture. Intel created a new platform; Vedicis unleashed its power for DPI and heavy broadband management. Combine the performance of Intel CPU that is evolving with Moore’s Law and the wealth of development proposed on Intel platform, and you have the ingredients of Vedicis DPI technology.

The quad-core servers were launched in 2009, the hexa-core in 2010. Incidentally, Intel sold its NPU business just in 2006…