DPI Announcements: NetLogic Adds a 100 Gbps Processor

 

NetLogic Microsystems announced the ".. XLP® II family of multi-core processors .. The XLP II processor family is designed to deliver over 100 Gigabits-per-second (Gbps) of network processing performance per device, and over 800Gbps in a clustered, fully-coherent system, which is an order-of-magnitude beyond anything currently available in the market .. a third-generation high-speed Fast Messaging Network® that provides higher-bandwidth, lower-latency communications among the 640 NXCPUs, and to support hundreds of billions of in-flight messages and packet descriptors among all the on-chip elements.  In addition to the Fast Messaging Network, the XLP II processors integrate a very advanced on-chip interconnect for the memory sub-system as well as a wide range of high-speed physical-layer and logical-layer networking interfaces .. The first members of the XLP II processor family will be available in the first quarter of 2012, with additional members expected to sample in the first half of 2012".

See "NetLogic Microsystems Unleashes Groundbreaking XLP® II, the World’s Most Powerful Multi-Core Communications Processors with Unparalleled Scalability to 640 NXCPUs™" - here.

DPI Announcements: NetLogic and EZchip Collaborate to Achieve IPv6 100G Performance

 

Two of the leading vendors of network processors, EZchip (here) and NetLogic Microsystems announced that ".. the companies are collaborating to deliver the industry’s highest performance, merchant, packet-processing solutions for IPv6-ready Terabit class systems.  By optimizing and implementing exclusive operational modes in both EZchip’s NP-4 100Gbps network processor (NPU) and NetLogic Microsystems’ NL11k knowledge-based processor, the companies are enabling customers to achieve enhanced performance and functionality when using both the processors together when compared to alternative solutions .. the growing requirement for deep-packet inspection throughout the network is driving an unprecedented need for knowledge-based processors with significantly higher performance and database capacity .. The companies have achieved broad design success across leading Tier One OEMs adopting the 100G NP-4 NPU and the industry-leading NL11k knowledge-based processor .. This powerful 100G combined solution from NetLogic Microsystems and EZchip is available immediately".

See "NetLogic Microsystems and EZchip Collaborate to Deliver High-Performance Packet Processing Solutions for Terabit Class Systems" - here.

DPI Announcements: Continuous Computing 200Gbps DPI Platform

   

Continuous Computing announced a new model for its ATCA based  FlexTCA platform, with 200 Gbps performance and 100GE interfaces (here), targeting network equipment vendors developing DPI products.

The highest performance we have seen so far for a standalone DPI product comes from Allot, with its recent 160Gbps announcement (See "Allot Boosts the Service Gateway to LTE" - here), also based on similar h/w components (ATCA, Netlogic packet processors). Nevertheless, Allot's Service Gateway is a fully functional product, while CCPU's platform still requires an additional application layer, which may or may not be able to explore the full performance potential of the platform.

While CCPU's new platform may seen as getting closer to an end-user solution (as it now combines DPI s/w, see below), Brian Wood (picture), CCPU's VP marketing, says that "CCPU draws a line just below the application layer and does not cross it"

"With this new platform family, network intelligence solutions such as Policy Charging and Enforcement Functions (PCEF) can exceed 200Gbps while wireless core and security applications can scale up to 320Gbps. Since “wire-speed” performance is so essential for DPI-based application differentiation and success, the cutting-edge FlexTCA family raises the ceiling on what is possible with mobile network packet processing platforms"

See "Continuous Computing Introduces Industry’s Highest Performance, Highest Capacity ATCA DPI Platforms" - here.

The new announcement includes the PP81 (here), a new packet processing blade with 40Gbps performance, a new switch (FM81, here) and a DPI software package (see "Continuous Computing Packages Qosmos, 6WIND and Wind River Packet Processing", here). With the FM81 "Continuous Computing’s new FlexTCA DPI platform is the first to come with native 100GbE interfaces".

DPI Announcements: NetLogic Multi-core Processor w/Layer 7 Processing

  

Embedding/integrating multi-core packet processors with DPI becomes more competitive as we now have 3 offering: Cavium (here) and Netronome (here) as well as NetLogic recently announced "..XLP316S multi-core, multi-threaded processor which integrates best-in-class NETL7™ Layer 7 knowledge-based processor technology.  This highly integrated processor that combines NetLogic Microsystems’ industry-leading XLP™ multi-core processing and NETL7 knowledge-based processing technologies is ideal for next-generation network security appliances, deep-packet inspection (DPI) gateways, intrusion prevention systems (IPS) and anti-malware gateways".  

See "NetLogic Microsystems Introduces Best-in-Class XLP™ Multi-Core Security Processor with Integrated NETL7™ Knowledge-based Processor" - here.

"The XLP316S .. is targeted to offer unparalleled performance of 20Gbps and 30 million packets-per-second (Mpps) .. The NETL7 knowledge-based processing technology features the Intelligent Finite Automaton (IFA) architecture which .. natively performs stateful cross-packet inspection in hardware. Furthermore, it integrates on-chip memory to provide ultra-low latency access to on-chip signature databases, allowing the processors to accelerate complex and iterative content inspection while eliminating the need to provision for the high-latency and costly off-chip memory that are required by competing solutions".

For comparison - Cavium claims for 40Gbps and Netronome for 100Gbps performance. Not sure if all claims are made on equal basis and assumptions ..

Chris O’Reilly (picture), VP of marketing at NetLogic Microsystems said “..As the industry’s only quad-issue, quad-threaded processor with integrated layer 7 knowledge-based processing technology, our XLP316S processor is uniquely suited to address the growing market for Layers 2 - 7 security processing in enterprise, small-and-midsize business (SMB) and carrier-class networks".

"Network Intelligence Alliance" Launched by Qosmos

      

Qosmos, a provider of DPI technology based in Paris, announced today "the launch of the Network Intelligence Alliance, consisting of technology providers involved in traffic capture, processing, decoding, analysis, leveraging of data as it crosses communications networks"

See "Qosmos: Network Intelligence Alliance Unites Vendors to Better Secure, Manage, and Monetize Network Economy" - here.

Founding members of the NI Alliance [site - here] include:

• Infrastructure suppliers for traffic capture and processing, namely, Continuous Computing, Napatech, NetLogic Microsystems, and Tilera
   
• Qosmos, the leading supplier of Network Intelligence technology
   
• Software vendors, namely, Click&DECiDE, Crossing-Tech, BreakingPoint Systems, and EdenWall Technologies, Rainstor, GfK Group

Brian Partridge, Vice President of Yankee Group's network research said. “We have been following the rise of network intelligence solutions and see their use growing in better securing, tracking and monetizing data across networks, devices, and users. Collaborative groups such as the NI Alliance will help to educate the market on this important area.”

Rumors: Cavium Received Proposal from Cisco - What does it Mean for DPI vendors?

  
Reported by Franklin Price, at benzinga site (August 31):

"Rumors circulating that Cavium Networks has received an unsolicited proposal from Cisco to acquire the company for $37.00 a share. Neither company could be reached for comment" (here).

Cavium (CAVM) closed at $25.21 on Thursday.

Cavium is a leading supplier of network processors for DPI (and other IP devices) equipment vendors. In Cisco hands, I wonder if they will continue to sell to other vendors - which are all Cisco's competitors. This will leave the market to its prime competitor, NetLogic/RMI.

"Cisco and the company’s manufacturing partners buy approximately 1% of all semiconductors produced globally, according to [Gary Mobley, an analyst at Benchmark report]. As a result, Cisco's chip vendors-such as Cavium and NetLogic-could be impacted. Cisco represented 25% of Cavium’s first-half 2010 revenue, up from 20% of 1H FY09 revenue. Cisco represented 29% Netlogic’s 1H FY10 revenue" (EETimes "Cisco's miss sends jitters in supply chain" - here).

See "DPI Announcements - Cavium Networks OCTEON II CN68XX" - here.

DPIBench - A Benchmark for Measuring DPI Product Throughput

 

EEMBC, Embedded Microprocessor Benchmark Consortium, announced "its intent to create a standardized, industry-endorsed method to evaluate the performance of Deep Packet Inspection (DPI) solutions".

 

See "EEMBC® Undertakes Design of Industry-Standard Network Security Performance Benchmarks" - here.

"In addition to creating the benchmarks that capture the performance of a system providing DPI, EEMBC will standardize the setup and certification methodology, as well as the standardized run and reporting rules."

“Most DPI product vendors use proprietary methods for measuring and specifying their products’ performance, which makes it very challenging for their customers to do an accurate competitive analysis,” said EEMBC president Markus Levy. “In the customer’s interest, we are encouraging all relevant companies, including the system manufacturers, network processor vendors, and networking and security application developers, to join us in this effort to ensure a meaningful and fair representation.”


While this effort targets security applications of DPI, I believe it may be relevant for traffic management applications as well.

Related posts:

• DPI Announcements - NetLogic 40Gbps Layer 7 Processing Solution  - here
• DPI Announcements - Cavium Networks OCTEON II CN68XX  - here

 

 

 

 

 

 

Ticonderoga Securities: "DPI will become critical in LTE networks"

 

In an article in Barron's Apurva Patel, Analyst at Ticonderoga Securities, selects his winners for "Next-Gen Video", following presentations in two industry conferences -the 10th annual Backhaul Strategies Conference  (here) and the 3rd annual Cable Next-Gen Video Strategies Conference - both organized by Light Reading.


See the report "Winners in Next-Gen Video" - here.

One of the conclusions is "We believe wireless- and video-service providers are facing some of the same challenges, mainly balancing revenue and higher costs per bit. In regards to telecom-service providers, we believe they are demanding higher bandwidth capacity with intelligent processing capabilities (including better security options), which we view as positive for semiconductor companies that enable equipment suppliers to meet their end-customer demands. For example, we believe security features, like Deep Packet Inspection (DPI), will become critical in long-term evolution networks.

I guess that DPI here relates more to the traffic management aspects (which is the main issue presented) rather than security (a valid aspect by itself). We saw a similar conclusion yesterday from Yankee Group (see "Yankee Group: 4G Video is a Killer App, but might be a Network Killer" - here).

The investment recommendation relate to semiconductor suppliers - including Cavium Networks  and NetLogic Microsystems (see "DPI Announcements - Cavium Networks OCTEON II CN68XX" - here) - anything beyond it - products, systems and solution vendors is not discussed.

DPI Announcements - Cavium Networks OCTEON II CN68XX

 
This time we go to the bottom of the food chain - the processing element. In recent years the computing power needed for packet processing grows faster than CPU evolution. As a result, vendors of smart traffic management devices need to explore new ways to increase performance and keep up with the demand for bandwidth over public networks.

In earlier days DPI devices were using "general purpose" CPUs (mainly Intel x86) but had to abandon the [easy to develop] technology to handle multiple Gbps links. Additional requirements including space and power consumption (see "Making Sense of Dense" by Tom Donnelly, EVP Marketing & Sales, Sandvine - here) led the leading vendors to use specialized network (or packet) processors or even build their own ASICS (application-specific integrated circuit).

See also - "Network Processor Overview" - here.

DPI vendors claim today for total performance of 40,80 or even more Gbps of traffic over multiple 10GE links.

The next step in the carrier deployment will include and upgrade to 40GE links - or even 100GE (see "40G vs. 100G optical technology battle will be decided by cost; 10G remains tough competitor" by Infonetics Research - here)

Cavium, one of the leading packet processor vendors, may have the answer. Cavium announced yesterday a new Multi-Core Processor  - the OCTEON II CN68XX described as family that "integrate 8 to 32 enhanced MIPS64 cores with up to 48GHz of 64 bit compute power in a single chip combined with over 85 L3-L7 application and security acceleration engines, virtualization features, 100Gbps of connectivity, and a revolutionary new Real Time Power Optimizer™ that dynamically adjusts power depending upon the application-level processing requirement.".

Press release - here

Cavium main competitor is RMI - now part of Netlogic.