Tuesday, October 28, 2014

Blueprint: Essential Elements of Transport SDN

by Stu Benington
Vice President, Cloud/SDN Business Unit, Coriant

There is a great opportunity emerging for service providers, driven by the widespread adoption of cloud-based applications by both enterprises and consumers. It’s a chance to capitalize on the elastic network needs of on-demand cloud-based services to create new sources of revenue and tighten control of CAPEX/OPEX costs.

The lack of elasticity in most of today’s static networks makes it difficult for service providers to deliver the on-demand bandwidth needed to support dynamic applications in a fast, cost-effective and profitable way. Also, the on-demand nature of cloud-based applications facilitates unpredictable traffic volumes that are often characterized by spikes in usage and variable traffic patterns. This challenges service providers to plan and engineer appropriate levels of network performance to satisfy users’ application demands.

In today’s fiercely competitive networking market, service providers need a new business model that enables them to capitalize on the proliferation of cloud-based applications and services. To this end, service providers must seek networking solutions built on Software Defined Networking (SDN) architecture.

An SDN-based solution transforms today’s rigid transport infrastructure into a highly programmable network that is as adaptive, efficient and scalable as data-center computing and storage resources. Transport networks become sufficiently flexible to deliver the bandwidth and performance needed for cloud-based applications on demand.

SDN shifts control of the network from internal elements to an external centralized controller and abstracts the underlying hardware infrastructure from the applications. The external controller now decides which network paths packets should take based on application requirements and the users’ willingness to pay. Then, using a standardized interface such as the NETCONF or OpenFlow protocols conveys its instructions to the forwarding plane, which remains within network elements such as routers and switches.

Transform the Network with Dynamic, Multi-Layer Transport

Service providers need to leverage the differentiated value of existing assets, including their network resources, networking expertise, and investments in the network brand. By deploying an SDN solution that combines dynamic multi-layer (L0-L3) transport with dynamic control, service providers can transform their network and service-delivery model while also maximizing return on their embedded investments. Such a solution enables the programmable network to respond to applications, rather than forcing applications to respond to the network. Using the intelligence in the external controller, the network can update dynamically, in real time, to accommodate the needs of a given application. In addition, these capabilities can be complemented by Network Functions Virtualization to move network resources to optimal locations that are then dynamically accessed by SDN.

The Essential Elements of SDN

A successful SDN has three essential attributes:

1. Programmability: Enables the network to adapt to the dynamic requirements of end users and applications, making it possible for service providers to introduce and modify services very quickly.

2. Multi-layer, multi-vendor end-to-end integration: As more applications and services move to the cloud, service providers must be able to integrate computing and storage resources across:

  • multiple vendors’ platforms
  • multiple segments of the network – access, metro and core; fixed mobile and data center
  • multiple optical and packet protocol layers

SDN helps manage this ecosystem of resources, including optical-layer advances such as flexible grid, ROADMs and photonic mesh, to deliver an end-to-end global view of the network.

3. Openness: Packet-optical transport networks comprise multiple vendors’ platforms and technologies. They depend on standards-based protocols for interoperability at both the virtual and physical layer. SDN’s open and collaborative software-based development process focuses on applications and accelerates innovation. By deploying a truly open SDN solution, service providers can introduce new functions and applications written by vendors, third-party developers, and service providers, with ease. It also provides flexibility for evolutionary steps from legacy environments to SDN programmability. An SDN solution that addresses these critical attributes is the foundation of the programmable network necessary to support and empower cloud-based applications.

SDN Use Cases


As SDN-enabled solutions move from laboratory trials to field trials, the hard evidence in support of the SDN business case is growing:
  • Bandwidth on Demand - An end user or an application may request temporary bandwidth between data centers, a lower latency on an existing connection (performance bursting) or an increase in bandwidth for cloud bursting. It can also provide user-defined connectivity for applications in the same “do it yourself” manner as using virtual machines (VMs).

    When the SDN-enabled network receives these requests, it determines the optimal path through the network. An optimal path is one that meets but does not exceed the requested performance criteria. For example, a service provider does not want to allocate a 10 Gb/s link if a 1 Gb/s link suffices. Similarly, a service provider does not want to allocate a 5-millisecond (ms) path to fulfill a request for a 10-ms delay if a 10-ms path is available. The goal is always to use the network’s least-cost resources first, with significant higher granularity than was previously available.

    Allocating network resources on demand benefits both the service provider and the customer. The service provider increases revenues by utilizing the network more efficiently and cost-effectively and more effectively monetizing network assets. The ability to create bandwidth interconnection services also increases network availability in terms of QoS, redundancy levels and protection. The customer gets the elastic bandwidth necessary to match their changing compute and storage needs, and also potential long- term cost savings.
  • Network Slicing - a process by which the service provider virtually apportions segments of the network across multiple products and layers to satisfy the specific requirements of a particular end user. For example, a service provider could use network slicing to partition services for a retail carrier in a wholesale environment or for a large enterprise. By eliminating the need to virtualize every single network element, the SDN-enabled network enables the service provider to deliver precisely what the end user needs – and do so much more easily, faster and cost-effectively than before.
  • Network Defragmentation - existing network connections are optimized by dynamically and continuously packing them in a more efficient manner on the network. Operating across both the optical and packet layers of the network, SDN-enabled network defragmentation packs flows to improve the network utilization. Because of SDN’s ability to provide a multi-layer, end-to-end view of the network, service providers can boost network utilization by 50- to over 90 percent.
  • Application-based Forwarding - SDN allows a service provider to set up the desired flow for an application according to that application’s specific requirements. For example, a service provider needs to upload into the data center all the data associated with an ERP system backup. Given the security and reliability requirements of that ERP system backup, it’s unlikely the service provider would choose the lowest-cost path to route that application through the network. Contrast that application’s requirements with those of a mobile user’s video application. When downloading that application, the user wants a lot of throughput and a low-cost connection, not necessarily a low-latency connection. SDN enables the service provider to take into consideration all the factors concerning a specific application’s requirements and come up with a recipe for the most appropriate and efficient flow through the network. Additionally, this service could be “calendared” at specific time intervals or done intermittently at the discretion of the provider.

SDN - A Framework for Sustainable Success

SDN is driving a network revolution. By transforming the transport network into a more elastic, adaptable and ultra-scalable entity, SDN creates a framework for service providers in which to create new, billable applications and services. That framework also encompasses the tools and techniques needed to create more sustainable business models and strengthen the ability to compete.
SDN is delivering significant financial and competitive benefits by enabling service providers to do the following:
  • respond to changing market conditions by creating new applications and launching new services faster
  • shorten dramatically the time required to dimension and provision the network resources necessary to support specific applications 
  • satisfy end-user requirements for on-demand bandwidth in an efficient and profitable manner 
  • reduce operational complexity through network simplification and automation 
  • reduce CAPEX and OPEX by a) enabling customers to tailor their application-specific connections and b) enabling service providers to distribute loads, with maximum speed and efficiency, among the most appropriate network resources 
  • free themselves from being locked in to specific vendors’ solutions and associated cycles of forced platform upgrades 
  • generate additional revenues that strengthen their margins 
By deploying the right SDN-enabled solution, service providers can capitalize on the opportunity to
migrate their multi-layer transport networks according to their individual business strategies and budgetary considerations. In doing so, they can protect their existing investments in switches/routers, optical transport platforms and network management systems in a programmable network that is essential for long-term success in an application-driven ecosystem.

About the Author

Stuart Benington is Vice President of the Cloud/SDN Business Unit at Coriant where he is responsible for leading the business unit, including strategy and R&D, focused on software defined networking (SDN), network virtualization, and cloud connectivity.

Prior to this role, Mr. Benington worked at Tellabs where he held a variety of strategy, marketing, engineering and product planning positions across several product groups, including Tellabs’ Data Products, Optical Networks, Network Management, and Managed Access Systems.  He started his career at Marconi (Reltec) working in product management for their broadband access solutions.

Mr. Benington has more than 20 years of experience in the telecom industry and holds Bachelor of Science degrees in economics and computer science from Purdue University and a Master of Business Administration degree from Northwestern University’s Kellogg Graduate School of Management.

One Minute Videos: The Third Network

The MEF recently outlined its vision for a new, THIRD Network, delivering Agile, Assured and Orchestrated services worldwide, where:
  • Agile means delivery of real-time on-demand services 
  • Assured means delivery of performance and security guarantees 
  • Orchestrated means automated delivery of services across multiple service providers.
This set of One Minute Videos introduces The Third Network. Links below:

Andrew McFadzen, Chairman of the MEF, discusses The Third Network and why carriers will be interested in setting a foundation for greater agility through multi-layer orchestration.


Nan Chen, President of the MEF, discusses The Third Network and its vision to deliver the best characteristics of the Internet in combination with service assurance of Carrier Ethernet 2.0.



Chris Purdy, Chief Technology Officer at CENX, explains that the First Network is the Internet, the Second Network is the telecom infrastructure and the Third Network will be the bridge between the two.




MEF GEN14

EMC Outlines its Vision for the Enterprise Hybrid Cloud

EMC outlined its vision for the Enterprise Hybrid Cloud, where an IT-as-a-Service model enables seamless interconnectivity between private data centers and public clouds, including VMware vCloud Air, Microsoft Azure, Amazon Web Services (AWS) and other EMC-powered cloud service providers.  The goal is to enable IT organizations IT to support all types of workloads and move them between cloud environments as needed.

The EMC Enterprise Hybrid Cloud Solution initially integrates hardware, software and services from EMC with VMware. It can be built on VCE Vblock Systems and EMC VSPEX reference architectures for accelerated deployment. The first release is EMC Enterprise Hybrid Cloud Solution Federation Software-Defined Data Center Edition with VMware, which is now commercially available.

Next year, the company plans to extend public cloud support to include the Microsoft Cloud Platform as well as OpenStack.

http://www.emc.com/about/news/press/2014/20141028-01.htm


EMC Acquires 3 Start-ups: CloudScaling, Maginatics and Spanning

EMC announced the acquisition of three start-ups focused on enterprise cloud networking:

  • Cloudscaling, a start-up based in San Francisco, developing an OpenStack-powered cloud infrastructure system for enterprises. Cloudscaling's  Open Cloud System (OCS) provides an operating system to manage compute, storage and networking in the cloud.
  • Maginatics, a start-up based in Mountain View, California,  is a cloud provider offering a highly consistent global namespace accessible from any device or location, unlocking enterprise hybrid cloud choice and flexibility for EMC customers and partners through interfaces into a variety of private and public clouds. EMC said the addition of Maginatics extends its cloud data protection strategy by enabling unified data protection and management across disparate private, public and hybrid clouds. Maginatics technology also facilitates efficient data mobility across multiple clouds with data deduplication, WAN optimization, handling of large objects and multi-threading. EMC expects to integrate Maginatics technology with existing EMC data protection software, storage and services.
  • Spanning, a start-up based in Austin, Texas, that provides subscription-based backup and recovery for “born in the cloud” applications and data. Spanning solutions prevent business interruption due to data loss in Google Apps and Salesforce.com (a solution for Microsoft Office 365 will be available in the first half of 2015). EMC saud the combination of its data protection portfolio and Spanning’s services uniquely positions EMC to help users confidently deploy data protection solutions across all applications and workloads, regardless of where the data is created or where the applications reside. 


Financial terms were not disclosed.

"What we have here is a trifecta of customer choice for hybrid cloud environments. Each company offers unique technology for delivering cloud abstraction and flexibility for customers. Together with the new EMC Enterprise Hybrid Cloud olution, EMC customers and partners are better positioned than ever to capitalize on the promise of hybrid clouds and deliver the agility of IT-as-a-Service that their businesses require,” stated David Goulden, Chief Executive Officer of EMC Information Infrastructure.

http://www.emc.com/
http://www.cloudscaling.com
https://maginatics.com
http://spanning.com/

NEC Launches PFTAP SDN Controller for Monitoring

NEC Corporation of America (NEC) introduced its PF6800 TAP SDN appliance for automated aggregation of telephony packet flows on a network using an Open Network fabric. 
The new controller, which is included in the recently announced version 6 release of NEC’s  ProgrammableFlow Software-defined Networking Suite, leverages an Open Network fabric to provide flow aggregation and orchestration; scalable filtration; and enhanced network traffic visibility. 
NEC cited the following benefits of using its PF6800 TAP:
  • Redirecting automated aggregation of packet flows to security and analysis tools
  • Easily add traffic filters via user interface, enabling selected traffic to be redirected to any port for further analysis
  • Improved visibility of network flows on logical and physical networks to easily identify TAP source data
  • Reducing the load on packet brokers, intrusion detection systems and other tools, through the use of OpenFlow switches for aggregation and filtration
“We’ve found that network visibility is becoming a major priority for network operators, and scaling out the monitoring equipment is difficult and expensive,” said Don Clark, director, business development, NEC Corporation of America.
List pricing starts at $36,000 for the appliance.

Gigamon Offers Active Visibility for High-Volume Traffic

Gigamon is preparing to release enhancements to the Gigamon Visibility Fabric for high-volume traffic in both scale-out and scale-up architectures for next-generation business infrastructures. The solutions are aimed at the increasing levels of east/west data center traffic from virtualization and distributed applications, which is driving higher traffic volumes from the leaf to the spine, spurring migration to 40Gb and 100Gb network connectivity.

The upcoming new enhancements to the Gigamon Visibility Fabric include:


  • Clustering capabilities for the GigaVUE-TA1 Traffic Aggregation nodes, enabling multiple nodes at the edge to act as a single logical unit when combined with an intelligent GigaVUE-H Series node at the core, for seamless extension of visibility to the edge.
  • The addition of the industry's first 40Gb BiDi TAPs, offered in compact 1 RU and 3 RU enclosures, and 40Gb BiDi optics into the Gigamon portfolio. This solution is ideal for customers considering a cost-effective approach upgrade from 10Gb to 40Gb using BiDi.
  • The industry's first combination of multiple 100Gb and 10Gb ports on a single blade allowing a higher degree of network and tool port consolidation to be achieved. The new 2-port 100Gb + 8-port 10Gb combination blade will be available on the GigaVUE-HD4 and GigaVUE-HD8 platforms.
  • Gigamon's Fabric Manager, GigaVUE-FM, will offer fabric-wide reporting capabilities, summarized and customizable dashboards, enhanced reporting to visualize the most/least utilized network/tool ports and traffic maps across the Visibility Fabric, audit trail support for security compliance and the ability to export reports for offline review.
  • The industry's first intelligent traffic visibility solution to automatically discover the topology of a connected network running Cisco Discovery Protocol (CDP) or Link Layer Discovery Protocol (LLDP). This unique capability allows automatic trace-back of network interfaces detected to be at fault by the Visibility Fabric and drastically cuts down mean time to resolution.

"We're seeing an increase in distributed mission-critical applications driving more east/west traffic below the spine of next gen data centers as well as an increase in the desire to gain a more pervasive view of network and VM-VM activity," said Ananda Rajagopal, vice president of product line management at Gigamon. "Unlike repurposed Ethernet switches that can only offer narrow visibility with no traffic intelligence, our new Visibility Fabric capabilities address higher traffic volume, allowing the best combination of cost, scale, reach and intelligence to be simultaneously achieved by administrators seeking uniform and pervasive visibility across their physical and virtual infrastructure."

http://www.gigamon.com/gigamon-announces-industry-leading-active-visibility-solutions-for-high-volume-traffic-monitoring

Gigamon Announces New Solutions for SSL Visibility

Gigamon is preparing to release a new GigaSMART traffic intelligence application that provides visibility into SSL sessions, thereby allowing deeper insight into infrastructure blind spots to help expose hidden threats or performance issues.  The company will also release a number of enhancements to its ‘Active Visibility for Multi-Tiered Security’ architecture detailed earlier this year, as well as associated upgrades to its Fabric Manager, which will centrally administer all of these capabilities across the Visibility Fabric.

"SSL sessions are either invisible to security and performance management tools or degrade performance by up to 80% if they are processed by the tools,” said Ananda Rajagopal, vice president of product line management at Gigamon. “By delivering SSL decryption as a common service to security and performance management tools, the tools can return to full performance. Further, because SSL is at the heart of today’s enterprise infrastructure, endpoints and DMZ servers are potentially exposed to attacks without the right level of traffic visibility.”

Gigamon notes that its GigaSMART applications can be ‘service chained’ together so that multiple visibility services can be performed in combination to address the specific needs of a range of vertical markets.

http://www.gigamon.com/gigamon-announces-new-solutions-for-ssl-visibility-and-enhancements-for-multi-tiered-security-


Orbital Suffers Launch Failure of Antares Rocket

An Antares rocket from Orbital Sciences Corporation exploded upon liftoff from NASA’s Wallops Flight Facility in eastern Virginia.  The launch was a catastrophic failure with significant damage to the launch facility.  No personnel were injured.

The rocket was carrying a Cygnus cargo logistics spacecraft destined for the International Space Station (ISS).  The spacecraft carried 5,050 pounds (2,290 kilograms) of supplies, including science experiments, crew provisions, spare parts, etc. It was the fifth flight of the Antares rocket and fourth deployment of a Cygnus spacecraft.

The Antares rocket was also carrying 26 mini Dove satellites from Planet Labs of San Francisco.
http://www.nasa.gov/ntv

Moka5 Raises $16 Million for Container Infrastructure

Moka5, a start-up based in Redwood City, California, raised $16 million in Series D funding for its next generation container designed to manage, secure, and extend both cloud and enterprise Windows workspaces across platforms, networks, and devices.

Moka5 offers a centrally managed container infrastructure for computers as well as mobile devices, an approach now being adopted by other vendors in the space.

Investors include Khosla Ventures, Highland Capital Partners, and NGEN Partners.

“It’s no longer a question of ‘if’ containerized workspaces will form the new enterprise client computing architecture - it’s ‘when’,” said Dave Robbins, CEO, Moka5. “With the strong market affirmation we’ve received from our Vanguard and Warlord releases and the tremendous amount of interest we’re experiencing from Project SkyNet, the Moka5 Enterprise AnyWare platform is at the forefront of today’s IT transformation. This has been a banner year for Moka5 and we’re pleased that our investors are as excited as we are with our prospects.”

http://www.moka5.com

Chelsio to Extend RDMA Over Ethernet to Microsoft's Cloud

Chelsio Communications announced that its T5 Adapters RDMA over Ethernet solution will operate in Microsoft's Cloud Platform System (CPS), which lets customers operate a scalable cloud built from Azure.

Chelsio's T5 adapters selected for Microsoft's solution are RDMA-capable Unified Wire adapters that simultaneously offload iSCSI, FCoE, RDMA and sockets applications, along with traffic management and QoS. Chelsio said its iWARP RDMA is well suited for CPS due to its high performance and interoperability. iWARP RDMA provides routability, reliability and congestion control ,which allow it to operate over any standard Ethernet infrastructure, from clusters to wide area networks and clouds.

Microsoft's CPS networking infrastructure is comprised of an Ethernet based frontend fabric and an RDMA enabled storage backend fabric, which utilizes the SMB Direct protocol for high efficiency, high performance storage networking. The SMB Direct protocol enables high throughput, low latency and high efficiency storage, and unprecedented performance in virtual machine mobility. Native system software support for iWARP RDMA in Windows Server 2012 R2 simplifies storage and Virtual Machines management for IT administrators, with no network reconfiguration required.

"iWARP continues to be the preferred solution for RDMA over Ethernet combining robust high performance with maximum interoperability and scalability. T5 is the ideal solution for the cloud and storage applications. We are thrilled with the selection of Chelsio's adapters within the Microsoft CPS," said Kianoosh Naghshineh, CEO, Chelsio Communications.

Vijay Tewari, Principle Group Manager, Microsoft, explained: "When we set out to build a high performance and high efficiency storage network, we thoroughly evaluated the different RDMA alternatives, and determined that Chelsio's particularly robust and high performance implementation of iWARP RDMA provided the best fit for the Microsoft Cloud Platform System. The plug-and-play nature of iWARP over standard Ethernet enables our customers to fully realize the performance benefits of SMB Direct 3.0."

http://www.chelsio.com/

CenturyLink Appoints Aamir Hussain as CTO

CenturyLink appointed Aamir Hussain as executive vice president and chief technology officer, where he will be responsible for design and delivery of next generation products, services and technologies. He reports to Glen Post, CenturyLink's chief executive officer and president.

Hussain most recently worked in the Netherlands as Liberty Global's managing director and chief technology officer for Europe. Previously, he was senior vice president and chief technology officer at Covad Communications.

http://www.centurylink.com

See also