Tuesday, December 15, 2015

Blueprint: Network Capacity Planning and Bandwidth Management in the IoT Era

by Leon Adato, Head Geek, SolarWinds

Enterprise network architecture has certainly evolved; from flat networks where everything was interconnected, to hierarchical models with enhanced security and now to a borderless world. But the one network metric that has remained a priority despite these changes is bandwidth, and by extension the individual traffic flows that comprise it.

Many enterprises have treated bandwidth as the elephant in the room—they have awareness of it and the tools to monitor it, but don’t have nearly enough insight into its usage. They don’t have visibility into usage history. They can’t correlate past performance to future trends. They lack any way to get a breakdown from the “big number” (bandwidth usage) to the root of the problem.

This has to change because, quite frankly, bandwidth is running out, and has been for some time now. What’s more is that there are additional bandwidth-hogging trends starting to crest the horizon.

Where We’re At

Already, BYOD and cloud have added extra layers of complexity when it comes to managing the network. With users connecting externally obtained devices such as cell phones, tablets, smart watches and even personal personal computers to corporate environments on top of corporate-provided devices, we’ve seen an exponential increase in bandwidth consumption. Furthermore, as more infrastructure is moved to the cloud, the network connections needed on that offsite infrastructure have also grown in both number and criticality. As a result, we network administrators have been tasked with redesigning networking schemes to adapt to these changes.

For those of us who have been through this and lived to tell the tale, we know that the key to success in the era of network complexity is preparedness in the form of network capacity planning and bandwidth management. Having a plan to manage current bandwidth issues and regularly analyze utilization information will best set us up to stay ahead of future issues that may arise.

Where We’re Going

Speaking of, what might be considered the second—yet much more challenging—installment of the bandwidth-hogging BYOD trend is fast approaching. Enter another now (in)famous acronym—IoT, or the Internet of Things.

Yes, it’s true that soon, even your company’s toaster oven may be connected to the network, along with a host of other devices and appliances—it won’t just be a swarm of HVAC, lighting and security controls or intelligent shop-floor tools that will expect Internet access; delivery trucks, trailers, shipping containers, smart pallets with onboard GPS, inventory management routing, sort and delivery elements, scanners and sensors of every variety will become Internet “things” using network protocols and bandwidth in unexpected ways. With more network devices in play than ever before, there will be an explosion of network traffic to accommodate the massive data volume, resulting in a harder time regulating the network. While some of the bandwidth being used will of course remain strictly internal, at the end of the day it’s all competing traffic, and competing traffic at a volume we haven’t had to account for in the past.

On top of that, because IoT will fundamentally change the way we humans interact with our environments, the ensuing complexities won’t just be about device and bandwidth entitlement added to the fully burdened cost of each employee. Environments will respond to the presence of humans, and user context—person, authentication, location, traffic, application—will all need to flow seamlessly as people move across traditional IT boundaries. So, as another consequence of IoT, IT departments will need to work closer than ever with the CIO, as well as legal, HR and other business departments.

And of course, it will be left up to us network engineers to sort all of this out.

Getting a Grip on IoT Network Capacity Planning and Bandwidth Management

If you’re like me, it probably seems like you’ve just barely gotten BYOD under control. The good news is that we learned some valuable lessons during that (long) process that are very applicable to getting a grip on IoT network capacity planning and bandwidth issues, too.

First and foremost is the need to closely monitor traffic—and not just the raw volume of network traffic, but application traffic, too. When it comes to IoT, traditional approaches like NetFlow will still be valuable, but IoT traffic will be more about application awareness than simple traffic monitoring and management. Quality of service monitoring is also very important, keeping in mind that IoT device responsiveness will be more important than traditional bandwidth-consuming things like email. Paradoxically, latency and reachability will be top priorities over limiting traffic.

And it won’t be enough to simply be a data collection platform or even a metrics dashboard solution. This monitoring will need to analyze more data than ever before, and transform it into concise, useful information to help us troubleshoot bandwidth-related network performance problems. Providing the breadth and depth of information needed to support devices, applications and networks in the era of IoT can only be done with an end-to-end, comprehensive monitoring solution.

As part of capacity planning for IoT, it will also be important to get IP address management under control and get gear ready for IPv6, which is what most IoT devices will prefer. Traditionally, many of us have managed IP address infrastructure with manual processes, which is labor-intensive, time-consuming and error prone. In addition, it leads to decentralized, fragmented, and outdated data. A simple request for a single new IP assignment can result in many hours of work, complex coordination and the likelihood for errors and conflicts, which in turn, can lead to a plethora of network problems. Just imagine what this will all look like when innumerable IoT devices start requiring their own addresses.

Finally, automate, automate, automate. In the IoT era, automating as much network management as possible will be more important than ever. At a time when there will be more devices accessing the network than you can shake a stick at, automation solutions will help to more quickly correct issues as they arise and will provide immediate remediation to reduce response times, significantly reducing potential network downtime due to any number of IoT-related capacity and bandwidth issues.

While it may seem crazy to say so, network capacity planning and bandwidth management in the IoT era really does not need to be a daunting task—we’ve been down this road before with BYOD. It’s simply a matter of remembering what worked then, being aware of the subtle differences we’ll experience with IoT and planning for them.

About the Author 
Leon Adato is a Head Geek and technical evangelist at SolarWinds, and is a Cisco Certified Network Associate (CCNA), MCSE and SolarWinds Certified Professional (he was once a customer, after all). Before he was a SolarWinds Head Geek, Adato was a SolarWinds® user for over a decade. His expertise in IT began in 1989 and has led him through roles as a classroom instructor, courseware designer, desktop support tech, server support engineer, and software distribution expert. His career includes key roles at Rockwell Automation®, Nestle, PNC, and CardinalHealth providing server standardization, support, and network management and monitoring.

About SolarWinds 
SolarWinds (NYSE: SWI) provides powerful and affordable hybrid IT infrastructure management software to customers worldwide from Fortune 500® enterprises to small businesses, government agencies and educational institutions. We are committed to focusing exclusively on IT Pros, and strive to eliminate the complexity that they have been forced to accept from traditional enterprise software vendors. Regardless of where the IT asset or user sits, SolarWinds delivers products that are easy to find, buy, use, maintain and scale while providing the power to address all key areas of the infrastructure from on premises to the cloud. Our solutions are rooted in our deep connection to our user base, which interacts in our thwack online community to solve problems, share technology and best practices, and directly participate in our product development process. Learn more today at www.solarwinds.com.

Got an idea for a Blueprint column?  We welcome your ideas on next gen network architecture.
See our guidelines.

Bell Labs Study Quantifies Residential vCPE Gains

A new study from Bell Labs claims that communications service providers can dramatically reduce their operating costs by up to 40% by ‘virtualizing’ complex functions currently deployed on residential gateways into the network cloud.

A virtualized residential gateway (vRGW) moves functions like IP routing and Network Address Translation (NAT) into the cloud, along with centralized management and control. This lets service providers introduce a more simplified, bridged gateway model making it easier for the user to successfully install, operate and maintain their home network, without having to make service calls or home visits.

Bell Labs estimates that a vRGW can reduce the costs of service fulfilment, assurance and lifecycle management up to 40%. In addition to reducing operating cost and improving the customer experience, a virtualized gateway model allows new services to be introduced more rapidly and consistently across the installed base of residential gateway devices.

Key Facts:

  • The Bell Labs study reveals cost savings in the following categories:
  • Service fulfillment: 7-12% cost reduction - a simpler residential gateway, extended auto-installation capabilities and network-based service capabilities lead to faster turning up of upgrades or new services and fewer home visits  by technicians to address service activation or upgrade issues. Home visits can easily represent over 80% of service fulfillment cost.
  • Service assurance: 63-67% cost reduction - service provider data shows that 30-40% of trouble tickets are related to issues deeper in the network. These can be resolved better by virtualizing and centralizing these deeper functions in combination with home device management capabilities.
  • Life-cycle management: approximate 66% cost reduction - although life-cycle management costs are relatively small compared to fulfillment and assurance costs, enhanced service velocity and agility improve service innovation, time to market, and revenue.

In addition to boosting profitability, a virtualized residential gateway delivers structural improvements in customer experience, service velocity and operational agility.

Alcatel-Lucent’s virtualized residential gateway solution is supported by both the 7750 SR and the Virtualized Service Router product line, and complemented by the Motive vRGW Controller  and the 7368 ISAM CPE and ONT product lines. The company’s vRGW solution is available now, with more capabilities being introduced in first half 2016.

"While the operating cost savings are essential to sustain profitable growth, having a well thought out  VRGW architecture is a pre-requisite to enable service providers to seamlessly extend the home network into the Cloud,” said. “This is critical to operationalize the rapid introduction of new features and automate the delivery of new value added services for the home network, without having the necessity to upgrade the customer premises equipment," stated Enrique Hernandez-Valencia, Consulting Director, Bell Labs.


ALU Builds Out vRAN portfolio with Red Hat, Advantech, 6Wind

Alcatel-Lucent announced collaboration agreements with Red Hat, Advantech and 6WIND. The partnerships are aimed at accelerating the delivery of commercial virtualized radio access network (vRAN) products.

Alcatel-Lucent is trialling its near-commercial vRAN with leading service providers in North America, Asia and Europe using a platform that consists of 6WIND, Advantech and Red Hat hardware and software. In 2016 Alcatel-Lucent will expand these trials using the CloudBand NFV™ platform to onboard and manage applications.

Some key points:

  • Alcatel-Lucent and Red Hat are expanding their existing collaboration on CloudBand and Red Hat Enterprise Linux OpenStack Platform to jointly enhance NFV-based features and better address the needs of the telecommunications market with open source. This includes the integration of Red Hat Enterprise Linux OpenStack Platform,  and Red Hat Ceph Storage within Alcatel-Lucent’s vRAN offering.
  • Advantech is supplying its Packetarium XLc PAC-6009, a high-density carrier-grade blade server with switching infrastructure that supports all processing, storage and switching elements in a compact size.  Advantech’s PAC-6009 server is ‘NEBS compliant’ for the North American market and fits in a standard rack to simplify network introduction.
  • 6WIND will supply its 6WINDGate product offering a variety of networking and security protocols and features to allow Alcatel-Lucent to optimize the performance of its vRAN technology.
  • Alcatel-Lucent is also working with a number of service providers to develop innovative NFV-based radio access network technologies. Additionally the company demonstrated its progress with China Mobile, Telefónica and Intel at Mobile World Congress in 2015 and announced the industry’s first live field trial of vRAN technology with China Mobile in July. 

"We are leaders in vRAN technology and use a strategy of collaboration to integrate leading third-party technologies and deliver world-class products that meet the unique needs of our service provider customers. Working with Red Hat, Advantech and 6WIND we are accelerating the availability of our vRAN, giving operators the flexibility and efficiency they require on a large-scale to meet subscriber data demand into the future,” said Glenn Booth, vice-president of Wireless Portfolio Management and Strategy at Alcatel-Lucent.


Advantech Launches Packetarium XLc Carrier Grade Blade Server

Advantech launched its Packetarium XLc carrier-grade blade server designed for mobile edge computing, where greater scalability and elasticity are needed to respond to fluctuations in subscriber demand and the introduction of additional new services.

Advantech said its Packetarium XLc is the first commercial-off-the-shelf (COTS) telco-grade server of its class to extend Network Function Virtualization (NFV) beyond the core network to both edge and access equipment. The system scales compute performance over nine Intel Xeon processor blades. Higher processing densities and lower power footprints are achieved through dual Intel Xeon processor blades that provide up to 144 powerful Intel Xeon processor cores in a compact 6U platform with a reduced depth of 400mm. The system routes traffic through two redundant switches that connect to the dual-star backplane creating an internal network with no single point of failure. It integrates Advantech’s Advanced Platform Management and is designed to meet demanding industry standards requiring five 9’s availability and NEBS Level 3 compliance.

Dan Rodriguez, General Manager, Intel Communications Infrastructure Division, said “Advantech’s Packetarium XLc is a good example of how members of the Intel Network Builders ecosystem play a key role in providing operators with the tools they need to actually go out and deploy a scalable and flexible NFV infrastructure built on Intel Xeon processors”.

“The Packetarium XLc enabled by a rich middleware partner ecosystem provides a solid NFV Infrastructure (NFVI) to application developers and content providers enabling them with the cloud-computing capabilities and IT-style service environment they need at the edge of the mobile network so they can drive new revenue streams for operators.” said Peter Marek, Senior Director x86 Solutions, Advantech Networks and Communications Group. “In addition, the system addresses real network deployment needs in locations with high mobile subscriber density and where greenfield installations are not possible.”


HPE Delivers Visibility into the Virtual Switching Fabric

Hewlett Packard Enterprise (HPE) announced network management and visualization software for virtual networks

The new Network Node Manager i (NNMi) software provides visibility into network topologies, helping companies to maintain control and extends their insight into the virtual switching fabric of their entire network. The software gives network managers the visibility into virtualized devices and topologies to help them ensure that their devices are connected, configured and performing as expected. When a device fails, NNMi can analyze events associated with the failure and help to recommend action. NNMi also provides predictive information that helps identify potential failures before they occur.

“Network managers today are unable to see a complete picture of their virtualized devices in real time, which limits their ability to ensure compliance and engineer the network for optimal  performance” said Balaji Venkatraman, PhD, Director of Product Line Management, Hewlett Packard Enterprise. “We are the first in the industry to provide a broad suite of network management system software with integrated fault, performance, configuration and compliance management capability that enables customers to optimize workloads to maintain application performance and resilience.”


Palo Alto Networks VNF Runs on Mirantis OpenStack

Mirantis and Palo Alto Networks announced a joint partnership and availability of Palo Alto Networks next-generation security as a virtual network function (VNF) within Mirantis OpenStack, a production-grade OpenStack distribution. The combination protects applications from cyberthreats while taking advantage of the agility, cost savings, and innovation of the OpenStack cloud ecosystem.

“OpenStack cloud environments are an attractive choice for many organizations because it gives them flexibility to accelerate innovation by rapidly deploying emerging technologies such as network function virtualization (NFV). We have teamed with Mirantis to give organizations next-generation security as a virtual network function (VNF) needed to operationalize OpenStack and effectively prevent successful cyber breaches,” said Marc Benoit, vice president of technical business development, Palo Alto Networks.

“As enterprises transition to a cloud infrastructure model, they migrate from hardware-based network services to a software-based, virtualized approach. Integrating Mirantis OpenStack with Palo Alto Networks VM-Series next-generation firewall is a good example of helping accelerate and secure this transition,” said Kamesh Pemmaraju, Mirantis vice president of product marketing. “Our Unlocked partner ecosystem makes it easy for companies to build an OpenStack cloud infrastructure with best-of-breed components.”


Telefónica Enters Alliance with MTN

Spain’s Telefónica Group announced a strategic alliance with the South African headquartered MTN Group.

The companies have agreed to work together to capture synergies in a number of strategic areas. Many of the initial areas are aiming to improve both companies’ stand in regards to Enterprise customers, including services to multinational companies in each other’s footprint, collaboration in M2M and new digital products and services targeting the B2B segment. Additionally, the Groups will also cooperate in various strategic initiatives and exchange best practices to capture the new industry opportunities. Both companies will also enter in discussions on how to engage effectively in International Wholesale, Devices and Network/IT Procurement.

The Strategic Partnership is supported by the Telefónica Partners Program, an initiative launched by Telefónica in 2011, which includes other leading telcos covering a total of 35
markets in Europe, Latin America, Asia and Middle East.


Zonamérica Installs Cisco for Business Park

Zonamérica, a special business park in Uruguay, has implemented a cloud-based Cisco collaboration solution to provide services to its clients.

This is the first project in Uruguay to deliver cloud-based services build on the Cisco Unified Computing System (Cisco UCS). The solution built for Zonamérica provides computing, storage, memory and software via an auto-provisioning system to the 200 customers within the park.