Unlocking the potential of AdvancedTCA in Network Intelligence

Service providers need to leverage Network Intelligence technologies to survive in the new network landscape. Whether these technologies require 10G or 40G capabilities, AdvancedTCA offers the price/performance balance to fit the bill.

3The new network landscape requires Network Intelligence (NI) technologies that shape and route traffic for applications like policy management and new service creation. However, increasing bandwidth demands require that service providers make critical decisions about the hardware infrastructure of their core network in order to take advantage of deep NI insights. Operators must not only determine whether the hardware architecture of their next-generation networks require 10G, 40G, or beyond, they must also consider how to best prepare for the transition. This article will examine AdvancedTCA (ATCA) technology
advancements, architectural benefits, integration technologies, and product developments, and how network operators can leverage them in support of their NI efforts.

Mobile infrastructure is facing rapid increases stemming from smartphone use and the rising importance of mobile video. Deep Packet Inspection (DPI) is the main technology used for the traffic shaping and load balancing necessary to keep networks running efficiently in the face of next-generation bandwidth demands, but deploying such advanced packet processing becomes a significant balancing act when the number of access points and the complexity of traffic both increase. Now, more than ever, mobile operators must find a way to deliver a high-quality experience to end users while adopting an efficient approach to spectrum consumption.

The need for increased bandwidth, better policy management, and the creation of services that generate profit are leading telecom carriers and network operators to invest heavily in their infrastructure. Implementing Long Term Evolution (LTE) technology within existing 2G/3G networks makes it possible to access greater capacity while driving down the cost per bit within the Radio Access Network (RAN); however, mobile operators must also deliver similar increases in capacity and cost reductions throughout the network in order to achieve meaningful Return On Investment (ROI).

One option for providers looking to increase Average Revenue Per User (ARPU) and reduce cost per bit is ATCA, a scalable, carrier-grade platform already in use in the core network. Capable of 40 Gbps switching in some cases, operators can leverage the performance and flexibility of ATCA with DPI technology to meet new consumption demands, as well as gather network and subscriber intelligence - and do so at a reduced cost.

In-platform flexibility for advanced packet routing

In particular, the need to manage the increasing traffic and demand for online and data services through dynamic policy management requires a high-density solution. ATCA affords carriers the opportunity to track subscribers as they move between cells, analyze applications and usage models with DPI and analytics on the front end, and provide policy enforcement resource functions deployed in ATCA.

Typical 14-slot ATCA platforms provide the ability to combine network fabric, the control plane, and the data plane into the same physical platform. DPI elements can be deployed in an ATCA platform either via specialized packet processing nodes utilizing dedicated Network Processing Units (NPUs) like Cavium Octeon II or Netlogic XLR-based blades, or via general-purpose compute nodes dedicated to packet processing workloads, through Intel Data Plane Development Kit (Intel DPDK) implementations, for example.

21
Figure 1: The network fabric, control plane, and data plane can be combine in 14-slot ATCA systems like UNICOM Engineering's A-13000 R2 to facilitate DPI applications.

DPI elements are commonly used as a network ingress point in ATCA platforms to perform packet classification and routing of data before it is disseminated to other computing elements via ATCA's internal network fabric. This is a common approach for SMS/MMS messaging platforms and IP Multimedia Subsystem (IMS) media servers, as well as Session Border Controllers (SBCs). ATCA's flexible architecture allows DPI elements to scale based on the workloads expected for any given deployment.

Lowering costs with improved performance

Because it is a high-density solution, moving to ATCA as an NI infrastructure platform lowers the cost per session; and moving to 40G ATCA lowers it even further. Although hardware costs increase with newer solutions, the actual cost per session decreases due to the board capacity of ATCA and the increased performance of 40G systems. For example, a 10G ATCA system with 12x Octeon boards can support 108,000 video sessions at $0.93 per session. When using a high-density 40G system featuring 12x Octeon II boards and a high-throughput 40G switch, on the other hand, the system can support 320,000 video sessions, dropping the cost per session to $0.41. For companies already using ATCA, accessing this improved price-performance ratio is as simple as upgrading the necessary elements to components configured to 40G requirements.

In addition, ATCA delivers cost savings when compared with conventional network servers. Three ATCA platforms at 250 W per slot can generate 200 A (or 9,600 W) of DC power per frame, resulting in 36 revenue-generating slots. A network server, however, can only generate 16 revenue-generating servers while consuming 192 A of DC power per frame.

Using ATCA to satisfy network demands

Today's environment offers companies developing ATCA-based systems the option to choose a range of components, or to develop a pre-integrated platform. By choosing the right solution, companies can leverage ATCA's modular architecture to minimize development costs and time to market while developing a market-leading solution.

With a well-designed system, operators can scale to achieve the wirespeed performance essential for leveraging DPI, and individual blades with greater processing power can be applied to enable operators to extend DPI functionality to additional traffic streams. Utilizing a consultative approach with a platform provider like UNICOM Engineering, scalable platforms can be designed to meet both today's and tomorrow's network needs (Figure 2). The focus should remain on expandability, upgradability, and supportability. As technology evolves, the chosen solution should evolve with it to take advantage of the latest advances and decreases in cost per session, while increasing total capacity.

22
Figure 2: Utilizing professional design and engineering services, Network Service Providers (NSPs) can build scalable platforms to meet the needs of today's and tomorrow's networks.

Scaling into the future, with Intelligence

Effectively, NI is being used to make intelligent routing and load balancing decisions within the network. DPI-based NI offers the means to deliver the experience users demand, but the resulting bandwidth requirements simultaneously dictate the need to scale. Transitioning from 10G to 40G with ATCA provides scalability and increased capabilities, while also delivering greater NI for deeper insight into consumer behaviors and policy management.

Austin Hipes currently serves as Chief Technologist and VP of Engineering at UNICOM Engineering, a division of UNICOM Global. He was previously VP of Technology at NEI and Director of Technology at Alliance Systems. Austin studied Electrical Engineering at University of Texas at Dallas.

UNICOM Engineering

www.unicomengineering.com

sales@ unicomengineering.com