While there are substantial differences between IPTV and Internet video, they are commonly blended together and hyped as the next big driver for everything from advertising, entertainment and social communication to business solutions and more. However, there still are some lingering questions about whether the network is really ready to support these applications and services with the high quality expected by end users (or the consumer).

Let’s start with a high-level definition of both types of video. IPTV is a commercial service that delivers TV using an IP network, which in most cases today is not the public Internet. As IPTV is a paid-for service, which requires high QoS and very high bandwidth, it typically is delivered inside the service provider’s restricted network — also referred to as the “walled garden.” (See my recent blog on net neutrality, “Can the Internet Really Be Neutral?”). Internet video is all the other video that can be accessed using the public Internet. This typically is referred to as “over-the-top” content created and stored outside of the service provider’s network. Yet, it consumes massive amounts of capacity on the public Internet portion of the network and ultimately can lead to congestion and hence reduce the user’s (or consumer’s) service quality experience.

Comcast Corp. and AT&T Inc. are two of the first service providers experimenting with limiting download bandwidth primarily because of the growing number of video users. Content delivery networks (CDN), such as Akamai and Limelight, are in a quandary over how to deliver and charge for higher bandwidth content, such as HD television, over existing networks. This is likely why in Apple’s recent earnings call, Steve Jobs called the Internet-delivered television segment of the market a “hobby” that he expects to continue into 2009.

In addition to these business challenges, there are technical challenges as well. Recent testing done by the Multi-Service Forum (MSF) regarding IPTV shows there are still difficulties creating an interface between IPTV set-top boxes and the IMS infrastructure. A vast array of potential screen sizes requires video to be encoded and/or delivered at different bit rates for the same content. Pirated or illegal content needs to be monitored and removed from countless Web sites. There are issues with QoS implementations over ADSL2+ and VDSL networks that impact IPTV quality.

But beyond these challenges lies a tremendous opportunity. The demand for network video is increasing rapidly. According to Infonetics Research, the number of subscribers of IPTV services is expected to grow from roughly 23 million in 2008 to about 70 million in 2011. Add to this other video services such as mobile video, user-generated content (UGC) and Internet on-demand video. A report by Juniper Research (not associated with Juniper Networks) predicts the value of the UGC market will increase from nearly $1.1 billion in 2007 to more than $7.3 billion in 2013.

All this content will be delivered over multiservice networks, which will need to balance bandwidth constraints, QoS demands, and ever-expanding types of end devices used to view the content. As the major service providers already are discovering, they are forced to balance planning and offering their own video services against the requirement to carry video content from external sources. Further, as IPTV delivery networks grow in scale, the core of the network will need the ability to handle more multicasting. And finally, content will be delivered to several new types of IP-enabled devices, such as Ethernet-connected televisions being developed by vendors such as Samsung. These new TV sets will deliver larger and higher density pictures, which will put additional strain on the network.

It will be very interesting and exciting to see how the telecom industry will respond to these challenges and how service providers and equipment vendors will adapt the network infrastructure to take advantage of the business opportunities offered by IPTV and network-based video.

Luc Ceuppens is senior director, head of product marketing, High-End Systems Business unit at Juniper Networks (JNPR).

During the past few weeks, we have witnessed a dramatic shift toward enabling the mobile Web: Google (GOOG), Nokia (NOK) and Research in Motion (RIMM) have all released handsets to compete with the iPhone; Sprint (S) launched its WiMAX-based Xohm network; Virgin Mobile (VM) is releasing an EV-DO handset and service plans that include 50Mb of data; and recent data from the Global Mobile Suppliers Association shows 250 HSPDA networks committed in 105 countries and 70 HSUPA networks deployed or committed in nearly 50 countries.

In addition, a recent Nielsen report shows that mobile devices are rapidly becoming the predominant portal for Web access and entertainment, specifically in developing countries, where fewer people have laptops and televisions. In other words, in these countries, users are more likely to watch videos, play games, and download music via a mobile connection rather than the traditional networks. Even in developed countries, where most of us have multiple TVs and computers, many have quickly experienced the addiction of 3G network speeds, and it didn’t take long to slowly switch our daily tasks to our wireless devices instead of reaching for our laptops.

So what does this all mean for network infrastructure? The first part is obvious. Accommodating the growing number of high-speed broadband mobile devices requires upgrading the radio access network, which is limited not only by the number of connections it can support, but also by the total amount of bandwidth that it can deliver. Many people tend to underestimate the impact of the mobile Web on the over-the-air network, but maybe recent developments surrounding the iPhone give you an idea. On Sept. 17, the Toronto Globe and Mail gave this explanation as to why RIM’s BlackBerry Bold is available in 13 countries, but not the United States: “Both devices use the same next generation (3G) network technology, but it now appears AT&T (T) wasn’t prepared for the bandwidth-hogging Apple (AAPL) device and doesn’t have the resources to launch the Bold until it cures its iPhone issues.”

It is no secret that the iPhone is a bandwidth-intensive device. In fact, a single iPhone can eat up as much bandwidth as 5,000 simultaneous voice calls, and it is not unusual for an iPhone user to consume up to 1 gigabyte of data per month. If sales of 3G phones keep going strong, carriers will run out of spectrum by the end of the year. When that happens, users are dropped to slower speeds and the Web-browsing experience progressively worsens, resulting in unhappy customers.

In such a scenario, a service provider really can’t afford to add more 3G phones onto its network, especially a device that might be very popular like the Bold. This issue has actually served as a wake-up call for the entire industry to the point where there are now discussions happening about providers moving away from unlimited data plans. However, there seem to be some alternatives. RIM has a network operations center that formats Web content before it is forwarded to your BlackBerry. It seems to work as the typical BlackBerry user only uses about 20 MB per month (although that number will likely change once the 3G BlackBerry reaches the market in high volume). Other vendors are believed to be setting up similar operations as well, and sooner or later, Apple will have to follow suit, or carriers will effectively kill the attractiveness of their phones by charging tiers for data usage.

While the issue with iPhone and other 3G phones is currently over-the-air bandwidth, this problem will ultimately be resolved, and we’ll witness a second wave of 3G devices that will dump huge amounts of bandwidth onto the backhaul and core of the network. According to a report from Wireless Intelligence in Sept. 2008, nearly 55 million people worldwide were subscribing to mobile broadband services in 91 countries, and this number is expected to grow to more than 1 billion in the next four years. In addition, the very nature of mobile communications itself will complicate matters as traffic patterns will be unpredictable. Since it is uncertain where users will be when they request services, carriers will no longer be able to assume that they can constrain traffic or subscription services to a particular edge location, but rather have to design the network for peak load.

As the mobile market keeps evolving, the underlying wireline transport network will need to follow. As mobile broadband speeds increase and more content becomes available, more traffic will flow through the core of the network. As more users roam seamlessly on a ubiquitous network with their sleek and shiny web-enabled gadgets, networks will need to scale to handle more connections and security sessions.

And as the mobile market continues to deploy 3G and is already thinking about 4G and beyond, next-gen and next-next-gen cores will need to give bandwidth yet another boost. 100Gbs to the rescue!

Luc Ceuppens is senior director, head of product marketing, High-End Systems Business unit at Juniper Networks (JNPR).

Deep packet inspection, or DPI, received a lot of negative press recently as part of the controversy around the creation of user profiles used for targeted marketing, and wrongfully so.

It is a typical example of confusing technology and its applications. The core of the controversy around targeted marketing is not about the DPI technology, but about the disclosure of its use to collect data on the online behavior of individual consumers and their ability to opt out. While the privacy issues rightfully dominate the media, it is easy to overlook all the other applications that are enabled by DPI (and that don’t require profiling).

Enterprise IT managers have long known the benefits of DPI. The ability to look into a packet well beyond just the header information allowed them to identify and “kill” malicious programs inside their networks, giving them more robust means of protecting their resources rather than just relying on firewalls or IDS and IPS systems at the network edge.

More recently, DPI found its way into service provider networks to support a broader set of applications. When carriers have more visibility into the packets and can inspect layer 2 through layer 7 of the data going across their networks, they are able to accomplish a variety of new functions including making routing decisions, filtering packets, even measuring heuristics and passing data on to another device. Applications enabled by deep packet inspection include tiered service offerings, custom QoS implementations, policy enforcement and lawful intercept to name a few. In essence, deep packet inspection allows service providers to unlock more value from their network and monetize this asset in more creative ways.

With tiered service offerings, they can offer a variety of services to different customers and charge them based on the services they use. Once the service provider sets a policy that defines which applications a subscriber can access, DPI can be used to enforce that restriction in the network, regardless of the software that exists on the client device.

Custom QoS implementations and policy enforcement allow carriers to determine the applications and data types that are using resources on their devices. With that knowledge they can make priority decisions based on criteria other than just QoS bits in the header. DPI can allow things like limiting peer-to-peer access, ensuring priority to VoIP or video services, or restricting certain applications altogether based on policy.

DPI can also allow service providers to identify specified traffic and divert copies of it for lawful intercept. Unlike simply putting a tap point on a line and collecting everything that passes through, DPI allows a filtered method of lawful intercept resulting in intercepted traffic that is highly relevant.

The above examples clearly show that DPI has many beneficial uses and that it can play a crucial role in helping service providers unlock value form their most important asset, their network. Even the use of DPI for targeted marketing shouldn’t alarm consumers if clear and concise policies regarding its use are communicated and enforced, and as long as consumers have an unambiguous way to remove themselves from the monitoring process.

Luc Ceuppens is senior director, head of product marketing, High-End Systems Business unit at Juniper Networks(JNPR).