Telecom Platform Deployment Featured Article
Telecom Platform Deployment and What You Need to Know About Packet Processing: Part 1
By Jamie Epstein, TMCnet Web Editor
As the telecommunications world remains ever changing, it is crucial for Telecom Equipment Manufactures (TEMs) to remain at the forefront of each new technology shift from 4G wireless to the mobile Internet.
I recently had the chance to speak with Austin Hipes, vice president of technology at NEI (News - Alert), in a new podcast about packet processing and how it will transform telecom platform deployments in the future.
First, let’s go over some brief definitions. Packet processing refers to the work done on the traffic of a network itself (primarily a packet-based network such as TCPIP) and is broken into two pieces. One element is header processing, also known as stateful packet inspections (i.e., general routing functions such as network address translation, basic quality of service, etc.). The other element is payload processing, or deep packet inspection, which looks deeper into the traffic itself and its content to provide advanced features such as policy-based management, anti-virus protection, monitoring and recording.
“The reason this is important is because we do more and more with packet-based traffic on a network, with data like voice, video and also standard traffic such as FTP,” Hipes stated. “This means networks keep becoming more complex and the overhead to process and manage the actual transmission of the data keeps rising. These strains required more work and more powerful systems, which take processing power away from the actual processing of the data itself as the traffic transmission becomes more and more difficult.”
Deep packet inspection allows you to look deeper into the actual traffic and determine more about what is being done at an individual session layer, which provides information about an individual process between two points as opposed to just general traffic. Hipes added, “An example of this would be in stateful packet inspection, we would know that you are trying to use FTP to transfer a file. By using deep packet inspection, we could actually look more into that file and determine the content (e.g., a video file, binary data, encrypted data, etc.) to more accurately determine the level of priority that data merits. This also helps determine the possibility of any policy violation or possibly a security violation such as a virus, instead of simply documenting how the data t is being transmitted.”
DPI demands increased capability of hardware, and Hipes commented, “Because we are looking deeper into the packets themselves as they are transmitted across the network, DPI requires much more processing to be done on the actual packets as they pass through. As networks keep increasing in performance and speed, you must have powerful systems that can look at those packets in real time to keep from negatively impacting the network performance. “
Unified threat management, which generally combines things like firewalling, intrusion prevention, gateway antivirus, anti-spam, is a big user of deep packet inspection. In the telco space, you have session border controllers, which are commonly used between two VoIP providers to provide connectivity, security, quality of service and other media services between disparate networks. And in the mobile 4G space, application load balancers are also becoming more prevalent.
According to Hipes, “We have more and more pervasive smartphones that run mini-applications on load balancers, and often times these applications rely on Internet services to provide the content. Load balancers are an efficient way to make sure that as you are going over a 4G network, these applications are properly balanced out. Hence, service providers can make the most efficient use of their processing and bandwidth to provide the best quality of service for those applications.”
Moreover, packet processing servers are currently in a transitional phase as the industry moves from servers with dedicated network processing boards added to them to handle all of the upfront work. These servers have the ability to determine the need for additional work, and if it is required, then hand it off to the general purpose server. This second server has its own applications running and determines what else must be done to the packet before doing the work and then passing it back to the network processor to leave the system.
“As we move forward, there is a lot of upcoming technology around multi-core processors and some new API layers that are going to allow a general purpose server to do much faster packet processing. This will enable mini general purpose servers to be standalone packet processors, “ Hipes stated.
To hear the full podcast, click here.
Be sure to stay tuned to the Telecom Platform Deployment channel, exclusively on TMCnet, for part 2 of this series!
Jamie Epstein is a TMCnet Web Editor. Previously she interned at News 12 Long Island as a reporter's assistant. After working as an administrative assistant for a year, she joined TMC (News - Alert) as a Web editor for TMCnet. Jamie grew up on the North Shore of Long Island and holds a bachelor's degree in mass communication with a concentration in broadcasting from Five Towns College. To read more of her articles, please visit her columnist page.
Edited by Carrie Schmelkin