Wednesday, March 12, 2014

Blueprint: SDN and the Future of Carrier Networks

by Dave Jameson, Principal Architect, Fujitsu Network Communications

The world has seen rapid changes in technology in the last ten to twenty years that are historically unparalleled, particularly as it relates to mobile communications. As an example, in 1995 there were approximately 5 million cell phone subscribers in the US, less than 2 percent of the population. By 2012, according to CTIA, there were more than 326 million subscribers.  Of those, more than 123 million were smartphones. This paradigm shift has taken information from fixed devices, such as desktop computers, and made it available just about anywhere. With information being available anywhere in the hands of the individual users some have started to called this the "human centric network," as network demands are being driven by these individual, often mobile, users.

But this growth has also created greater bandwidth demands and in turn has taken its toll on the infrastructure that supports it. To meet these demands we’ve seen innovative approaches to extracting the most benefit from existing resources, extending their capabilities in real-time as needed.  Clouds, clusters and virtual machines are all forms of elastic compute platforms that have been used to support the ever growing human centric network.

But how does this virtualization of resources in the datacenter relate to SDN in the telecom carrier's network? Specifically how does SDN, designed for virtual orchestration of disparate computational resources, apply to transport networks? I would suggest that SDN is not only applicable to transport networks but a necessary requirement.

What is SDN?

The core concept behind SDN is that it decouples the control layer from the data layer. The control layer is the layer of the network that manages the network devices by means of signaling. The data layer, of course, is the layer where the actual traffic flows. By separating the two the control layer can use a different distribution model than the data layer.

The real power of SDN can be summed up in a single word - abstraction.  Instead of sending specific code to network devices, machines can talk to the controllers in generalized terms. And there are applications that run on top of the SDN network controller.

As seen in Figure 1 applications can be written and plugged-in to the SDN network controller. Using an interface, such as REST, the applications can make requests from the SDN controller, which will return the results. The controller understands the construct of the network and can communicate requests down to the various network elements that are connected to it.

The southbound interface handles all of the communications with the network elements themselves. The type of southbound interface can take one of two forms. The first is a system which creates a more programmable network. That is to say that instead of just sending commands to the devices to tell them what to do SDN can actually reprogram the device to function differently.

The second type of southbound interface is a more traditional type that uses existing communication protocols to manage devices that are currently being deployed with TL1 and SNMP interfaces.
SDN has the ability to control disparate technologies, not just equipment from multiple vendors.

Networks are, of course, comprised of different devices to manage specific segments of the network. As seen in Figure 2 a wireless carrier will have wireless transmission equipment (including small cell fronthaul) with transport equipment to backhaul traffic to the data center. In the data center there will be routers, switches, servers and other devices.


Today at best these are under "swivel chair management" and at worst have multiple NOCs managing their respective segment. Not only does this add OpEx in terms of cost for staffing and equipment but additionally makes provisioning difficult and time consuming as each network section must, in a coordinated fashion, provision their part.

In an SDN architecture there is a layer that can sit above the controller layer called the orchestration layer and its job is to talk to multiple controllers.

Why do carriers need SDN?

As an example of how SDN can greatly simplify the provisioning of the network let's take a look at what it would take to modify the bandwidth shown in Figure 2. If there is an existing 100MB Ethernet connection from the data center to the fronthaul and it is decided that the connection needs to be 150MB, a coordinated effort needs to be put in place. One team must increase the bandwidth settings of the small cells, the transport team must increase bandwidth on the NEs, and routers and switches in the data center must be configured by yet another team.

Such adds, moves, and changes are time consuming in an ever changing world where dynamic bandwidth needs are no longer a negotiable item. What is truly needed is the ability to respond to this demand in a real time fashion where the bandwidth can be provisioned by one individual using the power of abstraction. The infrastructure must be enabled to move at a pace that is closer to the one click world we live in and SDN provides the framework required to do so.

SDN Applications

No discussion of SDN would be complete without examining the capabilities that SDN can bring through the mechanism of applications. There are many applications that can be used in an SDN network. Figure 4 shows a list of examples of applications and is broken down based on the type of application. This list is by no means meant to be exhaustive.


One example of an application that specifically applies to carrier networks is path computation or end to end provisioning. Over the years there have been many methods that have sought to provide a path computation engine (PCE), including embedding the PCE into the NEs, intermingling the control and data layers. But since the hardware on the NEs is limited, so the scale of the domain it manages is also limited. SDN overcomes this issue by the very nature of the hardware it runs on, specifically a server. Should the server become unable to manage the network due to size, additional capacity can be added by simply increasing the hardware (e.g. add a blade or hard drive). SDN also addresses the fact that not all systems will share common signaling protocols.  SDN mitigates this issue by not only being able to work with disparate protocols but by being able to manage systems that do not have embedded controllers.

Protection and Restoration

Another application that can be built is for protection and restoration. The PCE can find an alternative path dynamically based on failures in the network. In fact it can even find restoration paths when there are multiple failed links. The system can systematically search for the best possible restoration paths even as new links are added to the existing network. It can search and find the most efficient path as they become available.

SDN and OTN Applications

A prime example of SDN being used to configure services can be seen when it is applied to OTN. OTN is a technology that allows users to densely and efficiently pack different service types into fewer DWDM wavelengths. OTN can greatly benefit the network by optimizing transport but it does add some complexity that can be simplified by the use of SDN.

Network Optimization  

Another area where SDN can improve the utilization is by optimizing the network so that over time, it can make better use of network resources. Again, using the example of OTN, SDN can be used to reroute OTN paths to minimize latencies, reroute OTN paths to prepare for cutovers, and reroute OTN paths based on churn in demand.

NFV

In addition to applications, SDN becomes an enabler of Network Function Virtualization (NFV). NFV allows companies to provide services that currently run on dedicated hardware located on the end user's premises by moving the functionality to the network.

Conclusion

It is time for us to think of our network as being more than just a collection of transport hardware. We need to remember that we are building a human centric network that caters to a mobile generation who think nothing of going shopping while they are riding the bus to work or streaming a movie on the train.

SDN is capable of creating a programmable network by taking both next generation systems and existing infrastructure and making them substantially more dynamic. It does this by taking disparate systems and technologies and bringing them together under a common management system that can utilize them to their full potential. By using abstraction, SDN can simplify the software needed to deliver services and improve both the use of the network and shorten delivery times leading to greater revenue.

About the Author
Dave Jameson is Principal Architect, Network Management Solutions, at Fujitsu Network Communications, Inc.

Dave has more than 20 years experience working in the telecommunications industry, most of which has been spent working on network management solutions. Dave joined Fujitsu Network Communications in February of 2001 as a product planner for NETSMART® 1500, Fujitsu’s network management tool and has also served as its product manager. He currently works as a solutions architect specializing in network management. Prior to working for Fujitsu, Dave ran a network operations center for a local exchange carrier in the north eastern United States that deployed cutting edge data services. Dave attended Cedarville University and holds a US patent related to network management.

About Fujitsu Network Communications Inc.

Fujitsu Network Communications Inc., headquartered in Richardson, Texas, is an innovator in Connection-Oriented Ethernet and optical transport technologies. A market leader in packet optical networking solutions, WDM and SONET, Fujitsu offers a broad portfolio of multivendor network services as well as end-to-end solutions for design, implementation, migration, support and management of optical networks. For seven consecutive years Fujitsu has been named the U.S. photonics patent leader, and is the only major optical networking vendor to manufacture its own equipment in North America. Fujitsu has over 500,000 network elements deployed by major North American carriers across the US, Canada, Europe, and Asia. For more information, please see: http://us.fujitsu.com/telecom


)

Australia's NBN Co Tests Coriant's Terabit Super-channel

Australia's NBN Co has completed a trial of a one Terabit per second (Tbps) super-channel over a 1,066 km fibre optic ring in South East Queensland, Australia using Coriant's innovative FlexiGrid technology on the NBNCo Transit Network, which is a wholesale, open access network.

Coriant said this trial demonstrated a Tbps super-channel transmission in a 369GHz grid over 1,066 km. It showed a greater than 35 percent improvement in spectral efficiency which significantly increases the capacity of the system. The trial also demonstrated the flexible allocation of the super-channel by placing it in three separate locations within the c-band which maximizes use of existing fibre resources. This capability, in conjunction with the improvements in spectral efficiency, provides a maximum theoretical transmission capacity of 13 Tbps on existing hardware and fibre over a distance greater than 1,000 km.

NBN Co's Transit network is built using established optical fibre made available by third party carriers on long term leases and Dense Wavelength Division Multiplexing Equipment that provides connectivity from the access networks, including fibre and wireless, through to points of interconnect with NBN Co's wholesale customers.

The field trial was conducted on NBN Co's network using commercially available hardware and pre-commercial software, over the existing fibre that is currently being integrated into NBN Co's Transit Network.

Gary McLaren, NBN Co's Chief Technology Officer said, "We are pleased with the results of the trial with Coriant over our Transit Network. It highlights how established backbone infrastructure can be upgraded with sophisticated optical and electronic equipment to provide extra capacity for the future needs of the National Broadband Network.
"This proof point highlights that the existing transit network is robust and capable, as we rollout a mixture of Fibre-to-the-node technologies (FTTx) being designed to provide access to voice and broadband services faster, cheaper and more efficiently to Australian homes and businesses no matter their location across the country.
"As high-bandwidth applications and the growth of internet usage drive increased demand for network capacity, the ease of upgrading to higher transmission rates in our Transit Network will enable us to continue to deliver a reliable and high-quality broadband experience for our customers."

http://www.coriant.com

Telecom Italia Tests HSUPA 16QAM with NSN

Telecom Italia achieve uplink speeds of up to 11 Mbps using HSUPA 16QAM (High Speed Uplink Packet Access) technology in a lab environment.

The tests, which were conducted in Turin, used Nokia Solutions and Networks' advanced receivers, which can double the peak data rate and significantly increase the network capacity. The devices supporting HSUPA 16QAM are already available on the market.

NSN noted that the advanced Interference Cancellation receivers are already implemented in Telecom Italia’s network. They reduce interferences caused by high bit rate users, raising uplink throughput by up to 50% and extending the device’s battery life. The Frequency Domain Equalizer achieves an average of 10% to 20% gain in uplink throughput, and when combined with the HSUPA 16QAM solution, increase the peak uplink rate two-fold.

“We successfully tested NSN’s HSUPA 16QAM solution and achieved uplink transmission rates of up to 11 Mbps,” said Sandro Dionisi, director of Telecom Italia’s Lab. “The results prove that this technology is effective in increasing uplink data speeds which results in a superior smartphone experience for our broadband customers, especially when they post materials and updates online.”

http://nsn.com/news-events/press-room/press-releases/nsn-tests-hsupa-technology-in-collaboration-with-telecom-italia

Prolexic: High-Bandwidth NTP Amplification DDoS Attacks up 371% in 30 days

Prolexic Technologies, a division of Akamai that specialize in Distributed Denial of Service (DDoS) protection services, issued a high alert threat advisory on NTP amplification DDoS attacks.

Due to the availability of new DDoS toolkits that make it simple to generate high-bandwidth, high-volume attacks with just a handful of servers, Prolexic has seen a surge in this attack method. With the current batch of NTP amplification attack toolkits, malicious actors could launch 100 Gbps attacks - or larger - by leveraging just a few vulnerable NTP servers.

Some highlights of the threat advisory -- in just one month (February 2014 vs. January 2014):

  • The number of NTP amplification attacks increased 371.43 percent
  • Average peak DDoS attack bandwidth increased 217.97 percent
  • The average peak DDoS attack volume increased 807.48 percent

"During the month of February, we saw the use of NTP amplification attacks surge 371 percent against our client base," said Stuart Scholly, SVP/GM Security, Akamai Technologies. "In fact, the largest attacks we've seen on our network this year have all been NTP amplification attacks."

http://www.prolexic.com/


In December 2013, Akamai agreed to acquire Prolexic, a start-up based in Hollywood, Florida, for a net cash payment of approximately $370 million.

Prolexic offers a FIPS 140-2 SSL/TLS Layer 7 DDoS detection, monitoring and analysis solution for protecting data centers and enterprise IP applications from attacks.  Prolexic operates a DDoS "scrubbing center" in Ashburn, Virginia and San Jose, California as well as other facilities in London and Hong Kong.  The company says its solution was used to mitigate the largest Gbps attack faced to date (167 Gbps), as well as the world’s most powerful attack campaign (144 million packets per second). Its customers include some of the world’s largest banks and the leading companies in e-Commerce, SaaS, payment processing, travel/hospitality, gaming, energy and other at-risk industries. The company has previously disclosed global partnerships with HP, Level 3, BT, NTT and Datacraft.

Broadcom Intros 10/40/100G Lite-PHY for Data Centers

Broadcom introduced the first triple speed 10/40/100G Lite physical layer transceiver (Lite-PHY) designed for high-density data center applications.

The new BCM82322 Lite-PHY, which is fabricated in 28 nanometer (nm) CMOS, delivers the industry's highest port count density (12 full duplex ports) and reduces power up to 50% , consuming <150mw 10g="" duplex="" full="" nbsp="" p="" per="" port.="">
Key features:

  • Single 100GE CXP cPPI to CAUI Lite-PHY supporting SR10/CR10
  • Three 40GE QSFP+ XLPPI to XLAUI Lite-PHY supporting SR4/LR4/CR4
  • Twelve 10GE SFP+ SFI to XFI Lite-PHY supporting SR/LR/CR
  • Low-speed SFP+/QSFP+/CXP Data I/O
  • IEEE802.3ba 40GE CR4 Cl85 TX Training
  • High performance Adaptive Receive Equalization

"With the introduction of our latest 28nm triple speed Lite-PHY, Broadcom continues to demonstrate leadership in the PHY space and expand our industry-leading portfolio of 1/10/40/100G solutions," said Lorenzo Longo, Broadcom Vice President and General Manager of Physical Layer Products. "We remain committed to meeting the ongoing demands of our customers by delivering the highest level of performance while significantly reducing power consumption."

http://www.broadcom.com

Coriant Advances Collaboration with Juniper

Coriant reported further progress in its collaboration with Juniper Networks to create an integrated packet transport network (IPTN).

Recent joint R&D between the two networking innovators resulted in the first multi-vendor line side interworking between router and DWDM system suitable for long haul (LH) transmission. This IPTN solution is a combination of the Coriant hiT 7300 DWDM system, Coriant TransNet planning tool, and TNMS network management system with Juniper Networks MX Series and PTX Series routers.

The companies said their latest interoperability test was carried out on line side interworking of PTX 100G DWDM interfaces with hiT 7300 transponders and line system. The results highlight new use cases in seamless packet optical networks, which now include regeneration of router interfaces and handoff between routers from different vendors, which were not previously attuned on the line side.

"We see this collaboration between Coriant and Juniper as an important step that allows operators to offer reliable and scalable networks for richer services to its end-users by a tighter integration of the IP/MPLS and optical network layers," said Uwe Fischer, Chief Technology Officer at Coriant. "Both companies are highly motivated to create integrated packet network transport solutions to contend with the exponential growth in data and to offer long-term support to our customers from now and well into the future."

http://www.juniper.net
http://www.calient.com

See also