Procera Doubles NAVL Performance to 10Gbps/Core

Procera Networks announced "Network Application Visibility Library (NAVL) 4.3, the latest next-generation Deep Packet Inspection (DPI) software engine that provides real-time, Layer-7 classification of network traffic. NAVL 4.3 provides cutting-edge application identification and classification for today’s demanding consumer and enterprise Internet services, and offers a two-fold increase in performance .. NAVL 4.3 is generally available today for new and existing customers

..In recent benchmarking tests, NAVL achieved average performance gains of over 50 percent on standard x86 platforms, and doubled performance when utilizing an external connection management configuration. Leveraging an Intel Sandy Bridge processor Procera observed 10 Gbps of throughput per core and linear scalability as cores were added on a dual socket E5-2670 at 2.6 Ghz with HyperThreading and Turbo-Boost disabled. Integrators that require embedded DPI in specialized use cases like low-power devices (such as machine-to-machine [M2M] communication) or specialized CPUs (such as security) will especially benefit from NAVL 4.3’s global optimization. NAVL 4.3 also provides a simplified deployment process, consolidating hundreds of files down to one simple download".

Related post - "Procera: "lower than anticipated total addressable market" for NAVL" - here.


See "Procera Networks Releases NAVL 4.3, Doubling Performance for Embedded Internet and Application Intelligence" - here.

Juniper (W/Maxeler Technologies) Unveiled an Application & Packet Flow Accelerator

MaxelerLow Latency Platforms

Juniper Networks unveiled the "..QFX5100-AA, a new application acceleration switch, and the QFX-PFA, a new packet flow accelerator module, together providing a ground-breaking data center switch that consolidates compute resources and customizable logic into the network. 

Building on Juniper’s proven QFX product family, the QFX5100-AA and QFX-PFA in partnership with Maxeler Technologies’ customizable software logic will significantly accelerate business-critical applications in latency-sensitive computing environments.

Available in Q3 2015, the QFX5100-AA and QFX-PFA will enable financial services providers to streamline their network operations while empowering their customers to make split-second analysis based on significant amounts of data. Specific benefits will include:

• A holistic data center network architecture that builds on the Juniper Networks®QFX5100 with a new 40G Ethernet switch that combines the Intel® Xeon™ processor E3-1125C v2 with Broadcom switching silicon, and the QFX-PFA module based on the Altera multi-100G field-programmable gate array (FPGA).
   
• QFX5100-AA and QFX-PFA work with Maxeler Technologies’ dataflow programming environment for specific application programming that enables customers to develop and customize applications that best suit their requirements.
   
• Performance acceleration at the application level provided by using customizable logic to offload processing of compute-intensive, real-time, deterministic, latency sensitive applications"

See "Juniper Networks Advances Application Performance and Acceleration with New Compact Compute-Integrated Network Switch" - here and "Require high performance and application acceleration? Meet the new QFX5100-AA Switch" - here.

Sandvine States 155 Gbps NFV Performance (Beats Procera by 5 Gbps)

A post to Sandvine's blog by Don Bowman [pictured], the company's CTO, provides the results of a performance test to the vitrulaized version of its Policy Traffic Switch (see "Sandvine: 'leading edge CSPs around the world are currently working with Sandvine on virtualization pilot projects'"- here).

"Our Policy Traffic Switch (PTS) Virtual Series represents a particularly significant investment, as the PTS is the very visible, very physical hardware that many people picture when they think of Sandvine.

We are happy to report that our investments are paying off: in our latest performance testing, the PTS Virtual Series reached 155 Gbps (96.8% of a theoretical maximum of the 160 Gbps line rate). Importantly, this performance was achieved while the PTS Virtual Series dealt with real-world demands: using representative traffic mixes and requiring subscriber state awareness.

.. To test the PTS Virtual Series, we used a COTS server provided by Dell (Dell R730x, Dual Intel Xeon E5-2699 v3 (Haswell) processors, Cores 36/72 vCPUs, 128GB DDR3 ECC memory with 16x 10GbE and 4x 40GbE) and configured to achieve maximum performance based on readily available hardware components"

A month ago Procera announced that it has "..optimized PacketLogic/VTM virtualized software solution to demonstrate over 150Gbps running on Intel® Xeon® Processor E5 v3 Family-based commercial-off-the-shelf (COTS) platforms" (here)

See "The PTS Virtual Series: Virtualized Performance in the Real World" - here.

Qosmos: OEM w/Anritsu; Demos Intelligent Service Chaining w/Cavium; Joins WindRiver Ecosystem

Intense activity at Qosmos! The embedded DPI s/w vendor announced following:

• ".. it has joined the Wind River Titanium Cloud ecosystem, a program dedicated to accelerating the deployment of solutions for Network Functions Virtualization (NFV). The Qosmos Classifier [see "Qosmos Launches a Standards-based Service Classifier (2300+ Protocols and Metadata)" - here] has been pre-integrated and validated to fully benefit from the Wind River Titanium Server. This includes a straight alignment between the Intel® poll mode drivers (aka Intel DPDK®) along with the consistent networking (overlays support, etc.) support between the Wind River virtual switch and the Qosmos Classifier".
  
  See "Qosmos Joins Wind River Titanium Cloud Ecosystem to Accelerate Deployment of NFV Solutions" - here.

• "Anritsu has embedded Qosmos ixEngine® into its MasterClaw™ Multidimensional Assurance solution.Qosmos ixEngine provides real-time, detailed visibility of key mobile applications and services for telecom operators using Anritsu MasterClaw. They benefit from 24×7 visibility across their network for all subscribers and get unrivalled service assurance capabilities strengthened by Anritsu’s eoLive real-time experience dashboard and eoSight next generational analytics"
  
  Neil Tomlinson [pictured], COO, Anritsu said: “Thanks to Qosmos ixEngine, we can give our CSP customers a true picture of customer experience, as perceived by a subscriber, and also make a clear link between application usage and network performance”.
  
  See "Anritsu Enhances MasterClaw Service Assurance Solution with Qosmos ixEngine" - here. 

• It is "partnering with Cavium to showcase a demonstration of intelligent service chaining running on Cavium’s LiquidIO adapter. The demo shows a breakthrough combination of intelligence and optimized network processing power:

Qosmos Classifier

• Real-time intelligence is provided by a Qosmos-based service classifier, running on a Cavium LiquidIO adapter
   
• High performance is enabled by the Cavium LiquidIO adapter, which accelerates network throughput by offloading vSwitch processing from the x86 host

This new approach brings significant advantages for datacenter operators interested in dynamic, intelligent service chaining and optimized network infrastructure. They can programmatically create any type of service combination quickly and easily in a modern virtualized environment, while benefiting from extreme network performance.

Key technical features:

• Flow processing directly on the Cavium LiquidIO adapter for Layer 7 classification
• Open interfaces and integration with OpenStack and SDN controllers
• Provisioning of network rules consistent with the deployment of services
• Provisioning using RESTful interface using NETCONF/Yang models

See "Qosmos Showcases Intelligent Service Chaining in a Virtualized Environment Running on Cavium’s LiquidIO® Adapter" - here.

Telekom Austria Implements Connetem NFV-Based EPC w/Procera's DPI

Source: Vis's World premiere 

of VoLTE Service  (here)

Back in June Procera Networks announced that "Connetem Enhances its Virtual EPC w/Procera's DPI" (here). The product is now deployed by the "first operator worldwide to implement virtualization of the entire core network production". 

The two vendors, in conjunction with Telekom Austria Group’s Serbian subsidiary Vip mobile announced a "demonstration of the first fully virtualized LTE stack in a live network delivering data, voice and IMS services. Procera’s Network Application Visibility Library (NAVL) software engine provides the application identification, traffic visibility, intelligence and control for Connectem’s virtual Evolved Packet Core (vEPC) solution".

Two years ago, Procera deployed its h/w devices at Telekom Austria - see "Telekom Austria Selected Procera" - here]. Now, "Telekom Austria Group is now the first operator worldwide to implement virtualization of the entire core network production chain for mobile broadband services in a live environment. Instead of mobile traffic running on dedicated proprietary hardware and software, the data will now terminate in a general purpose IT data center running Intel x86 hardware and virtualization services. Additional solutions providers participating in the demonstration include Intel, VMware, Red Hat, Metaswitch, and Open Cloud".

Dejan Kastelic [pictured], director/CTO Vip mobile, said: “We are more than proud that Vip mobile achieved another successful NFV deployment. Being the youngest mobile operator and challenger in the market, we always seek for new opportunities to reshape the innovation landscape both in the local and regional markets. In this way we are directly improving our customers’ experience and expectation by introducing more efficient, cost-effective and energy saving solutions that enable faster time to market and flexible services delivery. Our dedication is to be the leader in providing new solutions to customers, vendors and local expert community by setting more advanced standards that will bring us closer to advanced European and world markets”.

See "Procera Networks’ NAVL Engine Powers Connectem vEPC for Telekom Austria Group" - here.

Qosmos Announces DPI for Virtual Switches (Intel first)

Qosmos continues with its NFV/SDN strategy (see "Qosmos Updates DPI Engine with SDN/NFV Readiness, Performance and New Protocols" - here) and announced the ".. new Qosmos Service Aware Module (SAM), which adds intelligence and service awareness to virtual switches. Virtual switch developers benefit from rapid time-to-market for solutions requiring built-in application intelligence, while telcos and enterprises profit from optimized network services and reduced infrastructure costs .. As a first implementation.

As a first implementation, Qosmos SAM™ is integrated with the Intel® DPDK vSwitch [see "Qosmos Integrates DPI with the Intel® Data Plane Development Kit" - here] [and] is available for pilot projects from Q1 2014 and will be commercially available from H2 2014".

"Qosmos SAM is integrated with a vSwitch to provide detailed, real-time traffic intelligence and enable service-awareness at the infrastructure level. It is based on Qosmos’ flagship product ixEngine, which is already established as the de facto industry-standard DPI engine for developers of telecoms and enterprise solutions. The new module classifies flows at the hypervisor level and the resulting protocol information and metadata are either used directly by the virtual switch or passed on to guest applications .. Qosmos SAM™ conveys real-time layer 4-7 intelligence via an open API designed to enable short development cycles. Qosmos is working with leading switch vendors and the open source community to drive innovation while ensuring API compatibility and improving performance".

See "Qosmos Announces SAM, Market’s First DPI Module for vSwitch, as the Foundation for Service-Aware Networks Based on SDN" - here.

[Heavy Reading]: DPI Vendors Shift Focus to SDN/NFV

A new report by Simon Stanley [pictured], research analyst, Heavy Reading finds that "Providers of DPI (deep packet inspection) technologies are shifting their focus to emerging software-defined network (SDN) and network functions virtualization (NFV) in an effort to boost network performance".

See related posts on Procera Networks (here), Allot Communications (here), Sandvine (here, here), Radware (here) and Qosmos (here)

Other findings:

• "SDN/NFV are driving a shift to virtualized platforms running on standard servers or dedicated hardware
   
• Strong ecosystem for DPI, policy management and analytics, including systems, software and processors
   
• Long Term Evolution (LTE) remains a leading driver for growth in policy management and analytics
    
• Consolidation in the market with Cisco, Intel [see "Intel Acquires DPI vendor Sensory Networks ($20M)" - here] and Procera [see " Procera Acquires Vineyard Networks for $28M" - here] acquiring key software suppliers
   
• Multicore processors from Broadcom, Cavium, Freescale, LSI and Tilera have been widely used in DPI systems. Network processors from EZchip and Netronome are used for pre-processing, load balancing and coarse DPI. Following the development of the Intel DPDK, x86 multicore processors and, in particular, the Intel Xeon E5-2600 family processors have been increasingly used for DPI and policy enforcement [see "New Intel Platform Provides 140 Gbps of DPI" - here]. The drive for virtualization will increase the use of x86-based platforms for DPI and policy management. The latest ARMv8-based multicore processors may also take some of this market. Multicore processors and network processors will be used to offload DPI and network interface functions, accelerating virtual platform performance".  

See "Virtualization Drive for DPI & Policy Management" - here and here.

New Intel Platform Provides 140 Gbps of DPI

Intel announced "..Highland Forest, a new communications platform that combines the Intel® Xeon® processor E5-2600 v2 with the next generation of the Intel® Communications Chipset Series 89xx. 

According to Steve Price [pictured], general manager of Intel’s Communication Infrastructure Division, Highland Forest will boost the performance of network traffic, delivering up to 255 million packets per second. Highland Forest will also create efficiencies for telecom customers though workload consolidation, streamlined R&D resources, common software tools and faster system integration"

According to an Intel "The Data Stack" blog post by Rose Schooler [pictured], GM and VP, Intel’s Data Center Group, (here) "With a dual-socket Intel® Xeon® processor E5-2600 v2 platform with the Intel Communications Chipset 89xx Series, performance scales up to 255 Mpps of L3 forwarding (64 bytes packets), 110 Gbps of IPsec throughput, 200 Gbps of OpenSSL* throughput, and 140 Gbps of deep packet inspection (DPI) throughput".

See "Chip Shot: Intel Announces Highland Forest, a New Platform that will Accelerate Network Transformation" - here.

Intel Acquires DPI vendor Sensory Networks ($20M)

John Ribeiro reports to networksasia that "Intel has signed a deal to acquire Sensory Networks (See - "DPI Announcements: Sensory Networks DPI S/W for Cavium Octeon II" - here), a provider of software pattern matching technology for network security applications. 

Chris Kraeuter [pictured], a spokesman at the chip maker, said he could confirm that Intel has signed an agreement to acquire Sensory Networks, but couldn't comment on the deal terms. Some reports said Intel paid about US$20 million for the company. Sensory Networks, with headquarters in Mountain View, California, also has a research and development office in Sydney, Australia".

See "Intel to acquire network security firm Sensory Networks" - here.

ipoque's DPI Engine Used for "Smart Pipe" Small Cell

ipoque announced a "new use case for its DPI to revolutionize consumers’ mobile Internet experience, as presented by Intel Lab’s engineers at Intel Developer Forum 2013 in San Francisco. Intel Lab’s Research Scientists showcased ipoque’s PACE engine, a Deep Packet Inspection (DPI) software library based on sophisticated detection techniques including behavioral and heuristic analysis, integrated into the “Smart Pipe” server used in small cells as a mobile Internet access gateway. The demoed prototype enables network operators to classify Internet traffic at the application level to ensure improved QoE for mobile Voice over IP (VoIP) calls and video streaming.

The “Smart Pipe” allocates all bandwidth available across all networks (WiFi or mobile networks) to these prioritized applications and also enables seamless handover for VoIP calls when connectivity is lost from one of the networks, ensuring the customer’s service is never interrupted.

See "New Deep Packet Inspection use case demoed at IDF13 helps small cell manufacturers to revolutionize consumer’s mobile Internet experience" - here.

Qosmos Integrates DPI with the Intel® Data Plane Development Kit

Qosmos continues to show progress in SDN/NFV based DPI (See "Qosmos Updates DPI Engine with SDN/NFV Readiness, Performance and New Protocols" - here).

The DPI vendor announced that it will "show a live demo of the world’s first virtual switch with embedded DPI at the Intel Developer Forum (IDF) .. The demo features Qosmos ixEngine® integrated with the Intel® Data Plane Development Kit (Intel® DPDK) vSwitch, where Intel® DPDK ensures high packet processing performance and Qosmos provides application intelligence on Intel Architecture"

See "Qosmos Announces Demo of World’s First Virtual Switch with Embedded DPI" - here.

[Infonetics]: SDP Systems ($6B Market by '17) Need API Exposure

A new report by Shira Levine [pictured], directing analyst for service enablement and subscriber intelligence, Infonetics Research finds that "Increasingly, operators are deploying service delivery platforms (SDPs) as a bridge between their existing and next-generation service efforts .. Business-to-business applications are growing in popularity, notably enterprise app stores and vertical-specific apps for telemedicine, smart grid, and telematics .. API exposure continues to be a hot topic in the SDP space, as operators look to capitalize on assets such as call control, payment systems, and subscriber information by providing secure access to both internal developers and third-party providers"

See also:

• [Infonetics]: $3.6B SDP Market Led by Oracle and Huawei (here) 
• [Analysys Mason]: SDP Market $4.6B for 2012; Ericsson Leads" (here)
• Infonetics 2012 forecast ($5.8B for 2016 vs. $6B for 2017 in 2013 report) - here.

"On the vendor front, Alcatel-Lucent and NSN have backed away from the SDP market, and M&A activity is heating up: Intel acquired Aepona and Mashery, and CA purchased Layer 7"

See "APIs playing key role in service delivery platform and M2M markets" - here.

Blue Coat Buys Security Analytics Vendor Solera Networks

Blue Coat Systems announced that it has "..entered into an agreement to acquire Solera Networks, the industry’s leading provider of big data security intelligence and analytics for advanced threat protection. The Solera DeepSee platform will add industry-leading security analytics and forensic capabilities to the Blue Coat product portfolio, delivering an end-to-end security solution that includes protection, remediation and governance and gives enterprises complete visibility into the content and context of advanced targeted attacks. The acquisition is expected to close within the next 30 days".

See also "DPI/Security/Analytics Market: Intel Invests $20M in Solera Networks" - here.

Solera DeepSee Dashboard

"Solera’s customers include organizations around the world in many market segments, including the Departments of Energy, Homeland Security and Defense, Hitachi, Qualcomm, Overstock.com, Parsons Corporation and Zions Bank. For these customers, Solera’s DeepSee platform helps defend against advanced persistent threats and targeted malware attacks".

See "Blue Coat to Acquire Solera Networks" - here.

[Intel]: What Happens in One "Internet Minute"?

Enlightening infographic by Intel, showing what's included in "Nearly 640K Gb of global IP data is transferred in just one Internet minute".

"So can our existing networks handle this explosion in network traffic and maintain consumer expectations for immediate access from multiple devices? And if the networks can expand to accommodate this growth, can they do it while maintaining security? Telco equipment manufacturers and service providers will be on the hook to ensure that we continue to enjoy access to information and entertainment on our mobile without any interruption to service".

See "What Happens in an Internet Minute?", by Krystal Temple (pictured), here.

DPI Announcements: Radisys Adds 40GE Controller for Intel DPDK; Uses Mellanox Technology

 

ConnectX®-3 EN Dual-Port
10 and 40 Gigabit Ethernet Adapters
with PCI Express 3.0 

Radisys Corporation announced a ".. single 40 Gigabit Ethernet (GbE) controller solution for the Intel® Data Plane Development Kit (Intel® DPDK). Radisys’ 40G solution enhances the Intel® DPDK beyond its current 10G functionality, delivering the high throughput. The 40G DPDK solution enables Telecom Equipment Manufacturers (TEMs) to significantly lower their cost-per-bit for Deep Packet Inspection (DPI) and security applications. In addition, by leveraging native 40G support, TEMs can reduce overall complexity and simplify their application development, shortening development time by up to 24 months".

"The Radisys 40G driver uses Mellanox ConnectX-3 NICs and is ported to both the Radisys ATCA-XE100 and ATCA 4600 compute blades"

Andrew Alleman (pictured), vice president, platform engineering, Radisys  said: "Our solution provides TEMs with a scalable solution with true 40G performance, making this an ideal solution for a variety of form factors from ATCA, Rackmount Servers and Network Appliances, and even to software"

See "Radisys Delivers Industry’s First 40g Solution For Intel Data Plane Development Kit" - here.

Universities Research Mobile Video Quality Measurement

   

Fredric Paul reported to ReadWriteWeb that "Intel, Cisco and Verizon are pouring $3.3 million into research at five prominent universities to improve video delivery over wireless networks. The first goal of the Video Aware Wireless Networks (VAWN) program? Find a good way to measure mobile video quality .. The research is important to finding ways to deliver the best video experience to the most people when the networks get overloaded; that is, minimize the problems that annoy people most"
 
"Because viewers perceive quality differently depending on what they're watching - sports versus talking heads, for example - quality isn't about throughput but experience .. Measuring subjective video quality is only the first step toward the larger goal of delivering high-quality video while boosting network capacity to handle the ever-increasing flood of data. The issue can’t be ignored: In 5 years an estimated 90% of Net traffic will be video, and 66% of mobile traffic will be video"
 
"The project is now in year two of its three-year, $1.1 million-per-year plan. By then end of the program, the VAWN team hopes to have a good method to measure perceived video quality, as well as algorithms and software tools to use that measurement to make network decisions".

 

 See "Engineers Struggle to Move Mobile Video Without Ruining It" - here.

Guavus: Characterization of OTT Traffic without DPI

 
Guavus announced it will "..demonstrate its Reflex Big Data Analytics platform, along with a suite of analytics applications tailored for cable operators' needs in customer experience management, network capacity planning and content & media optimization. At a glance, various use cases include:

• Characterization of OTT traffic without deep packet inspection (DPI)
• Subscribers' quality of experience via second-screen apps (iPad, Android and IP)
• Customer care interaction drivers (calls, tickets and truck rolls)
• Churn impact based on usage behavior, subscribed services and customer care interactions
• End to end visibility to content delivery and performance metrics

"The Guavus Reflex platform (chart below) is an end-to-end solution that provides actionable, contextually aware insights into customer behavior, network performance and application usage to enable timely, optimal decisions"

In February Guavus announced "..with growing traction in the US, Guavus currently counts three of the top four mobile operators and three of the top five Internet backbone operators as key accounts .. n addition, Guavus recently raised $25 million in a series-B growth round led by Intel Capital, Artiman Ventures and Sofinnovia Ventures" (here)

See "Guavus to Demonstrate Big Data Analytics Applications at The Cable Show" - here.

DPI/Security/Analytics Market: Intel Invests $20M in Solera Networks

 
Solera Networks announced ".. it has secured $20 million in Series D financing led by Intel Capital. Intel Capital is joined by existing investors .. round follows a year of tremendous momentum – with domestic and international growth exceeding 100% in 201 .. Its award-winning DeepSee platform continues to redefine rapid and proactive threat response using advanced Deep Packet Inspection capabilities to capture, index and classify all network traffic in real-time".

See video below about Solera Application Classification capabilities.

See "Solera Networks Raises $20 Million in Series D Funding Led by Intel Capital" - here.

Intel Demos DPI Enabling Technology

 

Intel addresses DPI developers. A post by Paul Stevens (pictures), Telecom Sector Marketing Director for Advantech, on Advantech NTG blog about the recent Intel Developer Forum, says:

"One key topic for discussion related to the latest innovations in processing and software technologies that are now being utilized to great effect in Deep Packet Inspection (DPI) applications. DPI is at the core of many new and emerging applications where multicore Intel processors offer a distinct advantage over their single core siblings. They maintain all the benefits of IA technology, including ease-of-programming, timely tick-tock refresh cycles and dependability. When coupled with optimized packet processing software, our Intel-based solutions enable the creation of a more scalable range of services at significantly higher bandwidths. At IDF we were able to demonstrate the improvements generated when you combine network acceleration and optimized packet processing software with Advantech’s Intel platform enhancements. Along with our software partners we’ll showed off a number of examples of the performance gains observed across a whole range of multi-core systems".

See "Intel Developer’s Forum 2011" - here.