In the search for more knowledge about the incredible pace of innovation and change that is driving major network transformation by enterprises and service providers; it is always a good idea to review the postings of those on the front lines. This is why the recent blog by Marten Hauville, Principal Solutions Architect (ANZ) for cloud networking specialist Alcatel-Lucent’s Nuage Networks business unit and Co-Organizer of the Australian OpenStack User Group, caught my attention.
Hauville in his blog raises and answers a timely question, “What’s up with the data center network?”
The reason this is so important is as Hauville notes, “We are in the midst of a transition in IT. Over the last couple of years the cloud has morphed from a disruptive technology on the periphery of IT into the mainstream.” In short, the world is going cloud and data center-centric.
Of the three pillars—Compute, Storage and Network— that are the foundation of the move to a data center-centric, software defined and controlled applications-based world, network historically has been a laggard when it comes to transitioning to next generation capabilities. However, as Hauville explains this is no longer the case. Indeed, thanks to Software-Defined Networking (SDN) and Network Functions Virtualization (NFV) the pace of innovation and adoption of cloud-centric transformations is accelerating. Hence, what’s up in regards to the data center network is so relevant.
Hauville starts with the assertion that: “Business competitive advantage these days is dictated by swiftness and agility, increasingly around business-driven applications that attain this advantage in the marketplace. This new edge is being pushed hard by enterprises that are adopting web-scale capabilities through software, drawing them into their inherent business products and practices.” He goes on to cite chapter and verse about how and why “Cloud IT” has become literally mission critical for enterprises in Australia and New Zealand.
Having made the case for Cloud IT, Hauville poses the question about how to enable the cloud to drive greater agility across the whole business. The answer is transforming the data center network. Yet, as he notes the Network presents some interesting challenges. In fact, the inability of Network to keep pace with Compute and Storage he says has led to a situation that, “Limits the overall efficiencies businesses could achieve from both their virtualization and initial private cloud investments.”
Cracking the network constraint challenge
What really caught my attention was the following statement by Hauville that: “This fundamental network constraint is not caused by the hardware capacities or bandwidth of the network. Far from it. The capacity and speed aspects of data centre networking have tracked well ahead of compute power with the availability and density of 10Gigabit, 40Gigabit and even 100Gigabit. The issue is due to limited evolution in the management, configuration and dynamism of these networks.”
I will not spoil why I have bookmarked the blog as a must reread reference. Hauville’s explanation of how the addition of next generation management, configuration, i.e., orchestration and control, can bring out the maximum value of all of the other technology upgrades taking place in data centers. He then goes on to make a very cogent case for Software Defined Networking (SDN) implementation as the means for achieving data center operational excellence.
Hauville closes with a caveat worth considering, “So if this future is set, and the underlying technology decision been made the key question now is not if you choose SDN but how you choose the right SDN implementation.”
Unfortunately, despite the embrace of traditional solutions of open source solutions for SDN, not all SDN solutions are alike. At an even higher level the caveat also should resonate since not all virtualization initiatives in general are not alike. The facts are that interoperability issues are going to be a major challenge for SDN. They are also going to be an issue for the NFV solutions that service providers are beginning implement. It will be fascinating to see how far and how fast the solution buyers will push the vendors to resolve these issues as internetworking and not just what goes on inside a data center or a federation of networked private cloud data centers comes to the fore.
Circling back to the question raised at the top about what’s up with data center networking, the answer is in two words, “a lot.” And, the caveat to this answer is that same as Hauville’s. Choosing the data center networking transformation technology that is right for your organization is a complicated challenge since there are options and vendors to be evaluated in the context of you unique requirements. However, such transformations are no longer about if but when, and because of the nature of how business is changing a sense of urgency about making the right move should be a driver.
]]>Currently, most route reflectors run either on a router that is dedicated to route reflection, or on routers that also perform other IP routing and service functions. Both scenarios have downsides.
Dedicated BGP route reflectors are a waste because route reflection functions require minimal data plane resources. Routers that juggle route reflection with other duties, on the other hand, may not have sufficient resources to support scalable route reflection.
Network virtualization offers a solution. A virtual route reflector, or “vRR” for short, can remove reliance on dedicated hardware and be adjusted up or down as needed through allocation of more or less resources to vRR virtual machines.
As a recent Alcatel-Lucent TechZine posting, Virtual route reflector delivers high performance, by Anthony Peres, Marketing Director, IP Routing portfolio, Alcatel-Lucent has noted, however, not all vRR solutions are created equal.
“Virtualizing an RR function is more than just compiling a software image to run on a virtualized x86 server,” noted the authors. “To meet the same level of stability and robustness that is offered today, virtualized network function implementations require a proven and stable software base optimized to operate within an x86 virtualized environment.”
A good vRR will take advantage of the multi-core support and significantly larger memory capacity of x86 servers. This can deliver a significant boost in performance and scalability for vRR.
“An implementation that supports parallel Symmetric Multi Processing helps unleash the power and performance of multi-core processing,” noted the blog. “This multi-threaded software approach offers concurrent scheduling and executes different processes on different processor cores. It significantly reduces route learning and route reflection times (route convergence times).”
The usefulness of vRR is not in question. But like many things, the devil is in the details.
]]>New Service Provider Revenue Monetization Model
Just as LTE has evolved to be the predominant technology for mobile broadband providers -- generating an average data volume per user of 168% higher than 3G data – so must the traditional charging model change. One charging model growing in popularity is Sponsored Data Charging.
Sponsored Data Charging enables mobile subscribers to view, stream, and benefit from sponsored content and use applications over the mobile service provider’s network without that data usage coming out of their monthly plan. The data charges that a subscriber would pay for the sponsored content are paid instead by the third-party provider owning the content. Even more importantly, a subscriber can test out certain sponsored applications and features for a short time to determine whether to subscribe, without impacting monthly data-plan limits.
Not only is Sponsored Data Charging built for massive broadband usage from all kinds of connected devices, it also can support emerging technologies such as VoLTE and NFV. Key target industries include advertising, retail, media, entertainment, healthcare, and financial services.
How Does Sponsored Data Charging Work?
Sponsored Data Charging requires flexible charging and rating models. One example is zero rating, where an application provider or third party pays for subscriber data consumed by the specific application or service, as well as data rewards, where the users are rewarded with additional buckets of data to extend their data plans.
For example, a mobile app developer could pay the mobile service provider a negotiated rate so the subscriber can use the app without worrying about data usage and overages. Zero rating data supported by advertising is a variation of this type of data plan. Another example is where a marketer provides a customer with a data reward as a bonus for making a mobile purchase, trying out their service, or for lead generation. Depending on the use cases with third parties, the real-time charging system needs to ensure accurate charges to the sponsor rather than to subscriber accounts.
Sponsored Data Monetization Partner Ecosystem
Sponsored data programs are proving to be an effective method of both monetizing mobile data consumption and attracting new subscribers on a global basis. To take advantage of this growing opportunity, Alcatel-Lucent is collaborating with Aquto. The collaboration takes advantage of Aquto’s deep understanding of the digital marketing ecosystem and combines with Alcatel-Lucent’s policy and charging expertise and experience.
Using the Alcatel-Lucent SurePay® platform, Aquto enables mobile service providers to rapidly roll out data sponsorship programs including data rewards and zero-rating of mobile apps/content. As shown in the diagram below, the service integrates with the existing IT infrastructure and with minimal investment, enables the service provider to monetize almost immediately by leveraging a global network of sponsors, which includes app publishers/developers, advertisers, and marketers.
Diagram: Sponsored Data Monetization Platform [Source: Aquto]
Aquto offers two programs for mobile service providers: Cloud-based sponsorships and the service provider user-engagement application.
Cloud-based sponsorships: To enable zero rating of mobile content, or offer data rewards to users, app publishers/developers, advertisers and marketers can leverage easy to integrate capabilities to keep users connected with their content and services.
Service Provider user-engagement application: In an increasingly competitive market, service providers have to devise new ways to keep users engaged. It is a destination to discover apps and services, offers from advertisers through which users can experience the benefits of zero rating and data rewards. The app draws users in and provides operators with the opportunity to engage users in a positive setting.
This sponsored data concept is not just about free Internet. It is quickly presenting itself as a vast opportunity for service providers to tap into a growing mobile ecosystem that is made up of app marketers and advertisers. Users are now spending more time with their mobile devices than in front of the TV or desktop. The timing is right for service providers to engage.
The combined expertise of Aquto and Alcatel-Lucent – including experience working with mobile operators -- has let leading mobile service providers rapidly roll out this new service that allows new revenue and monetization streams.
As they move into the cloud era, network operators need a service aware network operations tool to assure virtual network functions (VNF) management. They’ll need it to efficiently perform a variety of network operations tasks, including:
As described in a vEPC post related to converging NMS and VNF manager functions within the ETSI Management and Orchestration (MANO) architecture, operators need to evolve their network operations tools for NFV through tighter coupling the NMS and VNF manager functions. Specifically for VNF assurance, the blog states “Troubleshooting is simplified because traditional NMS faults/events are correlated with VNF related events/faults. The VNFM provides lifecycle management and automates the self-healing of VNFs.”
In addition to the ETSI MANO architecture, progress has been made in the ETSI specification for defining NFV Service Quality Metrics that strives to enable better engineering of VNF user service quality, more efficient fault localization and mitigation, and faster identification of true root cause of service impairment so proper corrective actions can be taken promptly.
As NFV service quality metrics and traditional network service performance are continuously monitored, a service aware infrastructure relationship model within a network operations tool will be important for it to be able to innately correlate events to the true root-cause of service impacting problems, without having to develop and pre-configure volumes of custom handling policy rules and scripts. In addition, this model will allow operators to perform a more rapid service impact assessment for network events under investigation, as well as speed fault isolation and resolution.
And to make this more advanced fault management meaningful for network operators, assurance visualization will help by providing intuitive views for easily understanding how a multitude of events and key quality indicators (KQIs) relate to each other, with clear visibility into the root-cause of problems. It will also insightfully give operators an understanding of the time-line for events and state changes in the network to give a better indication of cause and possible effects.
This blog is the 2nd in a series that discusses the evolution of network and service assurance. The 1st blog gives a general overview on how network operations tools can be more efficient.
ASSURING THE EVER-CHANGING STATE OF THE VIRTUAL NETWORK
VNF configurations will be far more dynamic than with physical network elements (PNF), presenting new challenges for network operations tools to keep pace with many events related to highly dynamic network state changes and elastic scaling.
Manual processes that piece together assurance data from disparate views will not be sufficient to keep pace in this highly dynamic NFV environment. And traditional real-time-only monitoring and assurance views will not be effective when a VNF could be here in 1 moment and scaled down and gone in the next. This means that there is a need for both current and historical events and state information to be intelligently processed with near real-time performance, and at large scale.
Consider how much more meaningful it would be for network operators if assurance views could be made more intuitive for easily understanding how all the network events and MANO related KQIs relate to each other. For example, wouldn’t it be more insightful for operators troubleshooting a service performance issue to have a timeline that shows the service impacting threshold crossing alerts (TCAs) as well as whether orchestration or network events occurred in the same general timeframe?
ENHANCING NFV ASSURANCE WITH SERVICE QUALITY METRICS
As VNF deployments increase, network operations tools will need to evolve with new NFV service quality metric definitions and provide intelligence for correlating the multitude of different events coming from the various types of NFV infrastructure and MANO elements. Specifically related to troubleshooting and root-cause analysis that works in coordination with VNF lifecycle management, operators need service aware visibility and traceability to the various possible service quality impacting layers.
For operations to be effective in a highly dynamic environment with network services that depend on both VNFs and PNFs for underlying network infrastructure, there must be a service aware understanding of the relationships between services and these VNFs and PNFs. And equally important, there also must be a mapping of how service quality events triggered by virtual machines, VNFs, and orchestration layers impact or trigger changes in dependent layers.
For example, when there are issues with virtual network provisioning latency or reliability or diversity compliance, these conditions may trigger actions within the orchestration layer. But as a primary concern of network operators:
Without a network operation tool that can provide this type of intelligence for assuring VNFs, operators will not have the visibility needed to understand whether a problem is within the scope of their control. And this is the type of information would not only be highly valuable for troubleshooting, but even more broadly for clarifying accountability for a localized problem across various organization groups from IT to the different network domain groups.
Operators require a unified network operations tool that has evolved with the intelligence to meet all of these new NFV related assurance challenges. This tool must possess a service aware model that is unified with NFV lifecycle management. It must scale and perform to keep pace with tracking huge volumes of events that reflect the continual state of flux of change across service quality impacting layers. (For more examples of service quality metrics that provide requirements for assuring virtual networks, please refer to the ETSI specification for defining NFV Service Quality Metrics.)
EVOLVING ASSURANCE WITH ADVANCED FAULT MANAGEMENT
Operators deploying NFV require advanced fault management that provides both current and historical visibility for root-cause analysis, so that active faults can be correlated with past ones as the state of the network changes. This historical fault correlation is essential for pinpointing the root cause of problems in the highly dynamic virtualized network where MANO triggered corrective actions could potentially make intermittently reoccurring customer impacting issues difficult to investigate.
And network and service assurance tools in the cloud /NFV era must scale to track the full history of related service impacting events so network operators can perform both real-time troubleshooting and trend analysis.
Tools also need to have the intelligence to detect reoccurring problems. Specifically, operators require a tool that can help them to assess whether corrective resolutions that were automated are successful, or whether they are failing. And if failing, whether the failures are persistent or intermittent, and whether there is an actionable probable cause against the network infrastructure within the scope of the network operator’s control. And amongst the high volumes of events, there will also be a need to suppress (or filter out) events that do not require an action by the network operations team.
The following video demo offers a deeper dive into an advanced fault management application from Alcatel-Lucent.
RELATED MATERIALS
]]>
Rarely does a video about network functions virtualization (NFV) captivate your attention like the one that Alcatel-Lucent recently uploaded about service innovation and lean operations in the context of NFV. Sometimes NFV can be a challenging concept to get your head around, but the video breaks it down with clear visuals and none of the PowerPoint that usually puts you to sleep.
If you haven’t seen the video you can watch the embedded version below. But also let me explain what the video is talking about.
Data center network operations are at the heart of telecommunications service delivery, but until recently nimble operations have been stalled by less than nimble infrastructure. New service creation required hardware deployments that both required up-front investment and limited service flexibility due to deployment times of days, weeks or even months.
NFV finally gets the network as lean and nimble as the virtual machines in the data center, allowing both the virtual servers and the network infrastructure to scale and change virtually as services are created or demand changes.
The Alcatel-Lucent video shows how companies can leverage NFV through the use of its CloudBand orchestration platform that manages network deployment and Nuage Networks’ network orchestration layer that does the network spin-up.
Through the use of network service chaining, which connects these services and also includes third-party infrastructure that works on the platform, operators can launch new services such as content filtering by simply clicking a few buttons for all the infrastructure components they want to spin up for the service, such as a WebRTC server.
The demo also shows how this NFV environment handles load variability and the hardware failure. When load rises, new virtual machines automatically spin up to meet the extra demand. When demand falls, virtual machines are reduced automatically.
The system also helps maintain high availability. In the demo video, when the operator needed to shift to a second data center after the first one failed, the platform automatically looked for where to set up a new backup to maintain high availability. This search included understanding the cost of service creation in various places, the weather factors, and other variables that network engineers usually need to consider when making a new deployment. Alcatel-Lucent calls this smart load placement.
Overall, the video is definitely worth a watch. Even if you already know a lot about NFV, seeing it in action is informative.
]]>From original TechZine article
Can the virtualized evolved packet core (vEPC) be deployed today in large scale, LTE networks? Mobile network operators (MNOs) are increasingly convinced that the vEPC has become viable both financially and technically. And I think so, too, based upon the advances made over the past year that I’ll discuss in this blog.
Advancements in vEPC scaling and performance
Early in 2014, the vEPC proofs of concept and field trials of virtualized mobility management and gateway products were limited in both scale and performance. But as the year progressed, advancements in the design and architecture used network functions virtualization (NFV) tools and capabilities that greatly improved their capacity and performance.
These improvements, together with other software enhancements, such as the Data Plane Development Kit (DPDK), have the vSGW/vPGW approaching the capacity and performance of dedicated hardware platforms.
Converged NMS/VNF manager: The key to seamless vEPC network operations
A lot of progress has been made with enhancements to the ETSI Management and Orchestration (MANO) architecture. However, rather than having separate element management system (EMS) and VNF manager (VNFM) functions, there’s been a move to converge these functions since both are integral to managing the VNFs. (The EMS described by MANO includes both network and element management (NMS/EMS) functions).
By unifying the VNF manager and NMS functions, an MNO can seamlessly manage and orchestrate the vEPC. This makes it easy for an MNO to perform VNF lifecycle management functions from the same NMS that is used on a day-to-day basis for network operations.
When EMS and VNFM are converged:
The traditional NMS Fault, Configuration, Accounting, Performance and Security (FCAPS) management function is now applicable to both the EPC VNFs and the physical network functions (PNF). This enables a common and consistent approach.
This also provides the topology and logical connectivity of the individual VNFs/PNFs and more advanced performance and SLA reporting. A single manager simplifies overall coordination and adaptation for configuration and event reporting between the virtualized infrastructure manager (VIM) and the NMS.
Troubleshooting is simplified because traditional NMS faults/events are correlated with VNF related events/faults. The VNFM provides lifecycle management and automates the self-healing of VNFs. It uses recipes to describe the vEPC VNF, its VNF components (underlying VM instances) and their interdependencies. Each VNF component has its own recipe, which includes a description of how to monitor, self-heal, and scale it.
With coordinated fault management and automated self-healing, the MNO’s operations team will have the visibility and intelligence to understand whether alarms are caused by normal maintenance activities or are indeed an emerging issue that they need to react to quickly. In addition, new advanced NMS approaches to network assurance visualization will speed problem assessment for both VNF and PNFs. These developments will also provide the VNF and network event data to support reporting and analysis.
When the VNFM and the NMS are combined into a single management functional instance, the management and orchestration of the vEPC VNF and integration of the vEPC into the existing OSS/BSS infrastructure is greatly simplified. This is because the VNFM/ NMS has complete knowledge and visibility of VNFs within the physical and virtual EPC network.
Is the vEPC ready for commercial deployment?
Based on the progress made in both the scalability and performance of the vEPC VNFs and the advances made in management and orchestration of the vEPC, 2015 will be the year for vEPC deployments to commence at some Tier 1 mobile operators. The momentum and confidence of mobile operators in NFV will make it a reality.
Alcatel-Lucent at Mobile World Congress
Alcatel-Lucent will have a large presence at Mobile World Congress in Barcelona. I will take part in a panel discussion on “Unifying Network IT and Telco IT” on Thursday, March 5th from 11.30 – 13.00.
We will also be demonstrating our vEPC at our booth. There you will be able to see the dynamic scaling of our Virtualized Mobile Gateway and the operational elegance of our NMS/VNFM system. I look forward to seeing you there and discussing how our vEPC solution can meet your NFV evolution plans.
Related Material
To contact the author or request additional information, please send an email to techzine.editor@alcatel-lucent.com.
]]>OpenStack isn’t an as-is solution for telco network functions virtualization (NFV) infrastructures. OpenStack is an open-source cloud management technology that provides many of the capabilities needed in any NFV environment. And this has prompted interest among many telco service providers.
But to realize the full benefits of NFV, service providers need NFV platforms that provide additional capabilities to support distributed clouds, enhanced network control, lifecycle management, and high performance data planes.
The OpenStack/NFV backstory
In 2010, RackSpace® and NASA jointly launched OpenStack®, an open-source cloud computing platform. Since then, the OpenStack community has gained tremendous momentum, with over 200 member companies.
Originally, OpenStack was not designed with carrier requirements in mind. So in 2012, a group of major telecommunication service providers founded an initiative to apply virtualization and cloud principles to the telecommunications domain.
The term network functions virtualization was coined for this initiative. Service providers called for vendors to build virtualized network functions (VNFs) and NFV platforms to help them become more agile in delivering services, and to reduce equipment and operational cost.
To address identified gaps in OpenStack and other relevant open source projects, major industry players established in September 2014 “Open Platform for NFV” as a Linux™ Foundation Collaborative Project. The intention is to create a carrier-grade, open source reference platform for NFV. Industry peers will build this platform together to evolve NFV and to ensure consistency, performance, and interoperability among multiple open source components.
There are 5 main areas in which OpenStack is currently lacking as a solution for telco NFV environments:
1. Distribution
In the IT world, enterprises want to consolidate their datacenters to reduce costs. But this is not always the best choice for NFV. Many NFV applications require a real-time response with low latency. NFV applications also need to be highly available and survive disasters. Service providers need the flexibility to deploy network functions in a distributed infrastructure — at the network core, metro area, access, and possibly even a customer’s premises.
Figure 1. Distributed NFV infrastructure
OpenStack supports Cells, Regions, and Availabilities Zones, but these concepts are not sufficient for the needs of NFV. Each OpenStack Region provides separate API endpoints, with no coordination between Regions. Typically, one or more Regions are located in one datacenter. The Cells component provides a single API endpoint that aggregates multiple regions.
With Cells, workload placement (“scheduling”) across cells is by explicit specification or by random selection. The Cells component doesn’t have a placement algorithm that is able to choose the best location based on the needs of the application.
The Horizon GUI is restricted to a single region at a time. There is no GUI able to show an aggregated view of the NFV cloud infrastructure. The OpenStack Glance virtual machine image manager is also limited to a single region. This means that the NFV operator would have to deploy images manually to the regions needed.
Bottom line: Service providers need a platform that will deal efficiently with the distributed NFV infrastructure necessary for low signal latencies and disaster resiliency. This infrastructure must also be manageable as a single distributed cloud with global views, statistics, and policies.
2. Networking
VNFs vary widely in their network demands. Because they are distributed throughout an NFV infrastructure, the baseline requirement for an NFV network is connectivity, both within datacenters and across WANs. Security dictates that different network functions should only be connected to each other if they need to exchange data, and the NFV control, data, and management traffic should be separated.
As network functions are decomposed – for example into data plane components and a centralized control plane component – network connectivity between these components needs to remain as highly reliable as traditional integrated architectures. Sufficient network resources should be available to ensure surging traffic from other applications cannot adversely affect NFV applications.
The network should be resilient against equipment failures and force majeure disasters. Latency and jitter requirements vary from hundreds of milliseconds for some control and management systems, to single digit milliseconds for mobile gateways and cloud radio access networks.
NFV networks will typically consist of a semi-static physical infrastructure, along with a much more dynamic overlay network layer to address the needs of VNFs. The overlay layer needs to respond quickly to factors such as changing service demands and new service deployments.
OpenStack Neutron is the OpenStack networking component offering abstractions, such as Layer 2 and Layer 3 networks, subnets, IP addresses, and virtual middleboxes. Neutron has a plugin-based architecture. Networking requests to Neutron are forwarded to the Neutron plugin installed to handle the specifics of the present network. Neutron is limited to a single space of network resources typically associated with an OpenStack region. It is unable to directly federate multiple network domains and manage WAN capabilities.
Bottom line: Service providers need a platform that will set up and manage local- and wide-area network (LAN and WAN) structures needed for carrier applications in a programmable manner
3. Automated lifecycle management
One of the greatest advantages of NFV as a software-based solution is its ability to automate operational processes. This includes the application lifecycle, from deployment to monitoring, scaling, healing and upgrading, all the way to phase out. Studies have shown that this automation will allow service providers to reduce operational expenses (OPEX) by more than 50 percent in some cases.
OpenStack Heat allows users to write templates to describe virtual applications (“stacks”) in terms of their component resources, such as virtual machines including nested stacks. Originally, Heat templates were based on AWS™ CloudFormation™, but more recently Heat Orchestration Templates (HOT) have been introduced that offer additional expressive power. Heat focuses on defining and deploying application stacks but does not explicitly support other lifecycle phases.
OpenStack Solum is a new project designed to make cloud services easier to consume and integrate into the development process. It is being designed to provide some of the missing lifecycle automation functions. There is some initial work on auto-scaling by combining the measurement capabilities of OpenStack Ceilometer with Heat. Heat is currently limited to a single OpenStack region.
Bottom line: Service providers need a platform that will automate not only deployment and scaling but also many other lifecycle operations of complex carrier applications with many component functions.
4. NFV infrastructure operations
The distribution of NFV infrastructures across many locations in a service provider’s network – as opposed to a few centralized locations – will pose specific challenges and impact the operational processes and support systems. NFV’s distributed infrastructure means that cloud nodes at different locations are added, upgraded, and/or removed more frequently than in a centralized cloud. These processes should be performed remotely whenever possible to avoid truck rolls across the coverage area.
OpenStack TripleO (OpenStack on OpenStack) is an experimental addition to the OpenStack family. The project aims at automating the installation, upgrade and operation of OpenStack clouds using OpenStack’s own cloud facilities. TripleO uses Heat to deploy an OpenStack instance on top of a bare-metal infrastructure.
Bottom line: Service providers need a platform specifically designed for a distributed NFV infrastructure, one that automates the complex software stack deployment and upgrade procedures.
5. High-performance data plane
Many carrier network functions (e.g., deep packet inspection, media gateways, session border controllers, and mobile core serving gateways and packet data network gateways) are currently implemented on special-purpose hardware to achieve high packet processing and input/output throughput. Running those functions on current off-the-shelf servers with current hypervisors can lead to a 10-fold performance degradation.
The industry is currently working on new technologies that have the potential to improve data plane performance on commercial off-the-shelf servers, in some cases to nearly the levels of special-purpose hardware.
Data plane performance, however, has been a fringe activity in the OpenStack community. Only recently, e.g., with the Juno release, more focus has been put on data plane acceleration. Juno offers support for requesting access for virtual machines to Intel®’s Single Root I/O Virtualization technology.
Bottom line: Service providers need a platform that will manage high-performance data plane network functions on commercial off-the-shelf servers.
Beyond OpenStack: What’s needed to make NFV work today?
Most service providers around the globe are looking for an open and multi-vendor NFV platform based on OpenStack. But as discussed, the OpenStack community is not strongly focused on some key NFV requirements. What’s missing is an NFV platform that goes beyond the scope of OpenStack to help customers realize reductions in CAPEX and OPEX, and improved service agility.
OpenStack is still under heavy development in many areas. As it matures, OpenStack will become more stable and richer in functionality, allowing it to better meet NFV requirements in certain areas. However, it is not expected to meet all requirements.
Service providers need a horizontal NFV platform that provides:
This approach will make it possible to break open today’s multiple application silos.
This article is based on the Alcatel-Lucent/Red Hat white paper CloudBand with OpenStack as NFV Platform.
To contact the author or request additional information, please send an email to techzine.editor@alcatel-lucent.
]]>Have you ever gotten your hands dirty and really implemented an NFV or SDN application? Six teams from academia and industry in Israel and Europe can answer with a resounding yes! These teams gathered in Haifa at the 4-day 2015 Winter School and Hackathon event, organized by Bell Labs, Alcatel-Lucent’s CloudBand team and Israel’s leading Institute of Technology, Technion. The event offered a full program to get acquainted with the fundamental concepts behind cloud computing, software defined networking (SDN) and network functions virtualization (NFV).
Eighty participants gained a clear understanding of enabling technologies, NFV and SDN challenges and barriers, and how to overcome the obstacles of implementing virtualized network functions in the cloud.
The program started with two days of in-depth technical lectures covering the principles of the cloud, server and network virtualization, OpenStack, and high performance packet processing for NFV among other topics. Following this, participants had the opportunity to get hands-on experience with CloudBand, an advanced NFV platform, learn how NFV changes operator roles and responsibilities, and how operational processes can be automated to reduce operational expenditure. One of the use cases shown was the automated deployment of an NFV application in a distributed NFV infrastructure.
After acquiring a solid foundation in the first three days, six teams took up the challenge to develop a real NFV solution. The task was to virtualize the DHCP function of a residential gateway. Virtualizing customer premises equipment, such as residential gateways, and moving some of their complex functions into the cloud has been identified as a promising strategy to reduce cost and increase service provider ability to quickly deploy new services.
The winner of the challenge was a team headed by Mladen Tomic from the University of Rijeka, Croatia, who implemented a solution that not only delivered the cloud based DHCP service, but was also capable of scaling to adapt to changing service traffic. Mladen said, “I pretty much enjoyed the whole event, from attending lectures on hot and interesting topics, exchanging ideas with other participants and having some great fun both learning and competing in the hackathon.” Congratulations to the winners and to all participants for their highly motivated participation!
The future of NFV will depend on a generation of students and engineers capable of grasping the opportunities and challenges of NFV, and we are convinced they will be the creators of advanced NFV solutions that we cannot imagine today. Anyone can join and create their own applications on a public version of the hackathon. VNF and NFV technology providers can also apply to participate in the CloudBand Ecosystem Program.
Over the past several years, I’ve met with many mobile network operators (MNOs) and discussed their plans for virtualizing the evolved packet core (EPC). It’s clear from the more recent conversations that MNOs are now convinced that the vEPC is both financially and technically viable for their networks. But is the vEPC ready for the MNO’s LTE consumer network? In this article, I’ll discuss why I now think that’s possible.
vEPC scaling and performance
Early in 2014, the vEPC proofs of concept and field trials of the Virtualized Mobility Management Entity (vMME) and Virtualized Serving Gateway (vSGW)/Virtualized Packed Data Gateway (vPGW) were limited in both scale and performance. But as the year progressed, advancements in the EPC Virtualized Network Function (VNF) design and architecture used Network Functions Virtualization (NFV) tools and capabilities that greatly improved their capacity and performance.
For control plane subscriber scaling, it is now possible to support up to millions of simultaneous attached users and hundreds of thousands of eNodeBs and small cells on a single vMME instance. This is comparable with today’s existing MMEs built on standard telecom hardware platforms.
In the data plane, the user capacity has increased significantly with the use of packet acceleration techniques. For example, Single Root – Input/Output Virtualization (SR-IOV) bypasses the hypervisor and enables Virtual Machines (VMs) to attach to the VNF (the vSGW/vPGW ) and share a single physical Network Interface Card (NIC) that functions as multiple virtualized NICsThis greatly improves speed and increases capacity by reducing processing overhead. These improvements, together with other software enhancements, such as the Data Plane Development Kit (DPDK), have the vSGW/vPGW approaching the capacity and performance of dedicated hardware platforms.
Converged NMS/VNF Manager: the key to seamless vEPC network operations
A lot of progress has been made with enhancements to the ETSI Management and Orchestration (MANO) architecture. However, rather than having separate element management system (EMS) and VNF Manager (VNFM) functions, there’s been a move to converge these functions since both are integral to managing the VNFs. (The EMS described by MANO includes both network and element management (NMS/EMS) functions). By unifying the VNF Manager and NMS functions, an MNO can seamlessly manage and orchestrate the vEPC. This makes it easy for an MNO to perform VNF lifecycle management functions from the same NMS that is used on a day-to-day basis for network operations. When EMS and VNFM are converged:
The traditional NMS Fault, Configuration, Accounting, Performance and Security (FCAPS) management function is now applicable to both the EPC VNFs and the physical network functions (PNF). This enables a common and consistent approach. This also provides the topology and logical connectivity of the individual VNFs/PNFs and more advanced performance and SLA reporting. A single manager simplifies overall coordination and adaptation for configuration and event reporting between the Virtualized Infrastructure Manager (VIM) and the NMS.
Troubleshooting is simplified because traditional NMS faults/events are correlated with VNF related events/faults. The VNFM provides lifecycle management and automates the self-healing of VNFs. It uses recipes to describe the vEPC VNF, its VNF components (underlying VM instances) and their interdependencies. Each VNF component has its own recipe, which includes a description of how to monitor, self-heal, and scale it. With coordinated fault management and automated self-healing, the MNO’s Operations team will have the visibility and intelligence to understand whether alarms are caused by normal maintenance activities or are indeed an emerging issue that they need to react to quickly. In addition, new advanced NMS approaches to network assurance visualization will speed problem assessment for both VNF and PNFs. These developments will also provide the VNF and network event data to support reporting and analysis.
When the VNFM and the NMS are combined into a single management functional instance, the management and orchestration of the vEPC VNF and integration of the vEPC into the existing OSS/BSS infrastructure is greatly simplified. This is because the VNFM/ NMS has complete knowledge and visibility of VNFs within the physical and virtual EPC network.
Is the vEPC ready for commercial deployment?
Based on the progress made in both the scalability and performance of the vEPC VNFs and the advances made in management and orchestration of the vEPC, 2015 will be the year for vEPC deployments to commence at some Tier 1 mobile operators. The momentum and confidence of mobile operators in NFV will make it a reality.
Alcatel-Lucent at Mobile World Congress
Alcatel-Lucent will have a large presence at Mobile World Congress in Barcelona. I will take part in a panel discussion on “Unifying Network IT and Telco IT” on Thursday, March 5th from 11.30 - 13.00.
We will also be demonstrating our vEPC at our booth. There you will be able to see the dynamic scaling of our Virtualized Mobile Gateway and the operational elegance of our NMS/VNFM system. I look forward to seeing you there and discussing how our vEPC solution can meet your NFV evolution plans.
Related Material
In doing preparations for what many have called “the circus” aka the annual Mobile World Congress (MWC)—which is arguably now the most important industry trade event as the world goes mobile and which takes place in Barcelona March 2-5—the excitement is already palpable. From all of the new cool devices of all shapes and sizes to interesting advances on things like antenna technology, Network Functions Virtualization, carrier aggregation, etc., the eye candy alone is almost overwhelming in terms of imagining the possibilities. However, MWC always is tantalizing because not only does it answer what we will see in terms of capabilities in the near-term, but also because of the questions it raises about the longer term.
In this regard I was struck by a recent blog by Michael Peeters, CTO, Alcatel-Lucent Wireless, entitled I think appropriately, The Circus is in Town. Peeters’ main point is summed up nicely where he says in characterizing his view on what’s the next big thing that will be the buzz of the show that, “...one thing is certain: its story will be about removing place and time constraints we took for granted.”
While I agree with where we are and where we are going to a certain extent in terms of removing place and time constraints, I happen to believe that the next big thing will be around what it has been and should be, i.e., utility. After all, as we move into always on/all ways connected broadband-enabled world, if all of the things Peeters points to about the Internet of Things (IoT), drones, wearables, more immersive customer experiences and the like are not both easy-to-use, trustworthy and useful, their monetization potential will not be maximized. And, let’s face it, the bottom line is the bottom line which is all about utility. This means utility as pervasively accessible and hopefully affordable and safe, as well as the perceived value we are willing to pay for personally and professionally that enables service providers to continue to accelerate the speed at which the future comes at us.
Talk of the town
If you think about MWC are more of a town meeting of all the stakeholders, along with being a good thing in the context of being a circus, which it is, Peeters’ observations about what will be discussed in terms of the short-term are spot on. In fact, his list of things that will be highlighted and discussed is worth review if you are lucky enough to go to Barcelona or not. They include his forecasts concerning what the show buzz will be about:
Applicability: LTE in new markets such as Public Safety. The ongoing deployment of VoLTE and potential further improvements. What needs to be added or changed to the existing technology?
Capacity: carrier aggregation in licensed bands, be it FDD or TDD, but also the combination of, and the tension between licensed and unlicensed spectrum and technologies. Do you go for LTE-U or Wifi-LTE aggregation?
Performance: specific deployment scenarios such as small cells for indoor and especially enterprise needs. Virtualization of the RAN. How and when do they make sense?
And, because the industry loves to discuss what comes next, despite 4G now finally being rolled out around the world, although “mature” might be a stretch given how far we have to go, 5G will be top of mind and interactions. Don’t get me wrong, this is a great thing. Who doesn’t like talking about the future? However, with things like VoLTE, Voice-over-Wi-Fi, Hotspot 2.0, IoT and M2M, really all in their nascent stages, my hope is that the industry is not getting ahead of itself. Indeed, the use cases that will emerge as to what is valuable as the mobile world moves toward being all IP and broadband are in most instances yet to be written.
This is a great thing as well. It is a so because it will be us figuring out the utility of new high-performance wireless networks and how to extract value from them and all of the new devices, business models (mobile payments for example) and competitive options that will determine what will be successful along with the who, what, when, why and how.
So as Peeters implies, by all means enjoy the show. Be entertained and enthralled. Like the circus, MWC is dazzling and if for no other reason can and should be appreciated for that alone. Indeed, take him up on the invitation to stop by the Alcatel-Lucent booth (Hall 3 – Booth 3K10) to learn more about the realities and possibilities.
In many ways we stand on the bottom of the on-ramp of the possibilities of the combination of pervasive computing and ubiquitous communications. The coming ability of networks to deliver more immersive experiences and better insights into how we can be more productive at work and enjoy more of the things that delight us in our personal lives makes this a unique time and a great time to be part of the buzz.
However, it is important to not lose sight amidst the technology of not just what it does but what it can and should deliver. In my mind the deliverables are utility and trust. It is my hope that at a high level these are the two things that are buzzing at MWC as they are what the industry should and can deliver as to what comes next.
Peeters has it right about the inexorable march of the industry toward providing us with the broadband infrastructure and agility the future is mandating in terms of breaking down the barriers of time and place, however, what this means in terms of deliverables creates fascinating open questions and opportunities and that is why the show is so engaging on so many levels. This may not a “new story,” but it certainly is an all important one.
]]>There’s no question that the network functions virtualization (NFV) technology around which many telecommunications carriers and vendors are rallying takes a page from the virtualization that already has taken hold in IT data centers. But you can’t judge a book by its cover. NFV and IT virtualization also have their differences.
One key difference is that while data center virtualization tends to rely on a centralized architecture, NFV calls for a distributed one, Andreas Lemke, marketing lead for the CloudBand NFV platform at Alcatel-Lucent, points out in a recent TechZine posting by Andreas Lemke, Marketing Lead, CloudBand NFV platform, Alcatel-Lucent titled, Why distribution is important in NFV.
“As the IT world virtualized, it found that a small number of warehouse-size data centers are more cost-effective than many small, widely spread ones. This is because companies that build data centers do not have to build and operate local access networks,” he wrote.
“In contrast to IT clouds, such as Amazon’s, distribution matters in NFV networks,” Lemke continued. “Many carrier applications have needs that are ill-suited to a centralized architecture.”
Those needs relate to availability, low latency (a key consideration in carriers’ radio access networks, where vRANs are being deployed), network offload (for which content delivery networks are being used), regulations, and security.
Consider network offload, for example. Because video and data have pushed ahead of voice as the most plentiful traffic on the network, there’s a need to optimize network operations for this more bandwidth-loving traffic. Using point-to-point video streaming in all cases is inefficient, notes Lemke, so carriers are leveraging content distribution and multicasting to make the most of their network resources, and a hierarchical, distributed architecture supports these network optimization efforts.
A distributed network also tends to equate to higher reliability and disaster survivability, he says, as when you have network resources in a broader geography the chances that all of those resources will be adversely affected by a man-made or natural disaster becomes lower.
And, while a distributed network creates more potential points of security risk, it also mitigates risk because more nodes exist, meaning there’s a greater chance parts of the network will be unaffected in the event of a security problem, and with the proper processes and tools, the attacks can be identified and isolated.
]]>
One of the things that will characterize 2015 is the trend that started picking up momentum in 2014 that communications service providers (CSPs) have developed a sense of urgency about transforming their networks. It used to be that if you were a network operator you could invest with some level of assurance that the hardware and the associated software to run it would be core to your network for possible decades before becoming obsolete. However, as everyone in the industry knows, this is no longer the case.
As the world becomes more software-centric in terms of service creation, delivery, agility, security and performance— to meet the tsunami of data heading operator’s way and to allow network operators to maintain their relevance as ecosystem hubs rather than “dumb pipe” providers—cost efficient and effective operational excellence and the need to be fast-to-market and fast in the market with innovative services and enhanced customer experiences have become paramount. It is why so much attention is being paid to thing like Software-Defined Networks (SDN) and Network Functions Virtualization (VFV).
The need for speed has become (pardon the turn of phrase) hyper-critical. However, with recognition of the need to transform and do so rapidly should also come the recognition that network operators cannot transform rapidly and successfully on their own. It may not “take a village” to get transformations in the fast lane and done right, but it certainly takes trusted partners. In fact, Olivier Gueret , Senior Marketing Manager Wireless Transmission at Alcatel-Lucent, in a recent TechZine article, Rely on partners for your network transformation, makes a nice case as to the vital role partners can play in helping develop and expedite successful network transformations.
In fact, Gueret explains why professional services in particular are important in network transformation projects for a variety of reasons including filling in skills gaps and having experience in all of the complexities of such projects. After all, network transformations from my own observations are like trying to change jet engines while a plane is at 30,000 feet. They are extremely complicated, especially since every customer is unique, and the plane needs to stay in the air and perform at optimal levels even as parts are replaced. There is also interesting challenges regarding the costs of change and how to quantify that the ends justify the means.
Gueret in his posting posits the case made above, i.e., network transformation is no longer a nicety it is a necessity. He goes on to highlight that this really is a case of different strokes for different folks. In fact, he points to a recent Ovum study that when it comes to the reasons to transform operators are divided in two camps:
As he notes, both approaches have the same goals of transforming their network to increase revenues and reduce OPEX, they certainly diverge as to how. This leads to falling into some traps which reliance on a trusted partner with deep network transformation expertise can help mitigate.
Gueret points out the hidden costs of “home-made” network transformations which can translate into additional costs. Cited additional costs from going it by yourself if you are an operator include: costs of unexpected delays caused by poor planning and sequencing; costs from poor quality assessments of infrastructure capabilities; and costs from over-dimensioning, .e.g., spending on things that will not be used or cannot be optimized.
The case for relying on a trusted partner
As Gueret details, the case for relying on a trusted professional services partner is a compelling one. He notes that such a partner, “Can define, plan and execute a transformation efficiently, even if most operators have in-house competencies to do it themselves.”
The benefits he cites are:
The article goes on to point out how professional services are part of a broader set of capabilities for upgrading network infrastructure, and that partnering on a variety of fronts can enable operators, regardless of where they are coming from, to shift their business models. This means relying on a variety of trusted partners to not only to prepare and execute their network transformation but also to manage and maintain their networks.
This would let operators shift their business model to focus on their core activity: managing their commercial offers and their customers. This is a reality that is summed up well in the chart below from the posting.
Figure 3. Enabling operators focus on customer facing activities
The message is a powerful one. The urgency is there for operators to transform their networks for a host of well known reasons relating to operations costs and competitive necessities, and despite a cultural history to do almost everything themselves, network operators by relying on the expertise of others have the opportunity to meet their cost objectives and concentrate on what they do best. This means not just listening to the voice of the customer but hearing them and reacting quickly in ways that encourage loyalty and the willingness to trust the operator when evaluating the purchase of new products and services.
]]>Their networks, which traditionally have been based on turnkey network elements running software on purpose-built hardware, are moving to a software-centric model. In this model the true value lies in the software, while the hardware is typically of the commercial-off-the-shelf variety.
Network Functions Virtualization (NFV) is the name of this new architecture, which not only embraces the model of instituting network functionality in software and running it on industry-servers, but also allows applications and services to leverage those resources whenever and wherever they are.
The success of virtualization in the data center has demonstrated the power of running network capabilities on virtual machines. That’s powerful because it allows networks to be more fluid so they can meet shifting demands. It’s also powerful because it can result in cost savings, given less – and less specialized – hardware is required, and given virtualized environments (in which one server can host various network elements) tend to consume less power than environments featuring a collection of appliances.
NFV also can help facilities-based network operators effectively reinvent themselves to be more agile, so they can better compete with faster and often smaller over-the-the-top service providers.
Reducing equipment costs and power consumption, and expediting the introduction of new services and features were among the key goals laid out by ETSI’s NFV group, which got the network functions virtualization movement rolling a couple years ago. Founders of the NFV group within the European standards body included AT&T, BT Group, Deutsche Telekom, Orange, Telecom Italia, Telefonica, and Verizon.
Network operators that want to get started with NFV, suggests Andreas Lemke, marketing lead of the CloudBand NFV platform at Alcatel-Lucent, should take advantage of what he describes as “5 must-have attributes of an NFV platform.” These include:
Finally, and as important as all of the technology, Lemke says that those wishing to get started with NFV should select partners that can provide the same five 9s reliability, quality of service, and security in the new virtualized environment as they enjoy with their existing networks.
There is a growing industry consensus that NFV will become the architecture of the future for networks that are agile, applications friendly, high-performance, interoperable and secure. In fact, not only is there consensus but there is traction in the market for NFV solutions as service providers look to transform themselves to be as accommodating as possible in a profitable manner to the dynamics of rapidly changing market requirements. However, not all NFV solutions are alike which is why the Lemke attributes list is one worth consideration as part of an NFV evaluation.
]]>
It feels like it was just a few months ago when you could read articles in the trade press lumping together SDN and NFV with NFV being a form of SDN or vice versa. Yes, both somehow are about virtualization and about converting hardware into software. Today – after numerous proofs-of-concept run by service provides around the globe – we know the role of SDN as virtually indispensable for NFV solutions that aspire to deliver the kind of agility and operational simplification we all expect from NFV. Only SDN can deliver quickly enough the (virtual) networks needed for newly deployed network functions. Alcatel-Lucent has recently demonstrated a complete virtual evolved packet core (vEPC) including a virtual IMS/VoLTE deployed in less than 30 minutes.
NFV and SDN enable on-demand service composition by steering traffic through a sequence of middle-box service functions (service function chaining), such as firewalls and traffic optimization. For example, an enterprise or consumer customer can use a self-service portal to check off the desired functions, which causes virtual network functions to be deployed or scaled and (per-subscriber) routing policies to be changed automatically (flow-through provisioning).
Likewise, NFV responds to changing traffic within minutes by spinning up additional virtual machines within the same data center but also in a data center close to where the traffic demand originates. NFV enables rapid software upgrades while containing the risk of service degradation. We are even seeing demand on the horizon for adopting Devops models in the telco domain.
A classical operational model with change requests being sent to the networking department is no longer up to the task. The network needs to be as dynamic as the server infrastructure and it is clear that only SDN can fill the bill. This will be a stepwise process and not just any SDN will be suitable for NFV. Telco networks are not only about dropping packets in on one side and the packets popping back out at their destinations. Telco networks are designed to deliver enough capacity, high enough performance, security and high availability for the critical services running over them in an end-to-end geo-distributed environment.
Clearly, SDN is right for NFV but it needs to be the right SDN. Read the white paper “The right SDN is right for NFV” to learn about critical network requirements for NFV, SDN use cases and four stages of SDN integration into NFV bringing different degrees of reward to service providers. Alcatel-Lucent CloudBand™ and Nuage Networks® VSP are discussed as an example integrated SDN/NFV solution.
]]>The advantages to mobile operators of network functions virtualization (NFV) and moving to a virtualized evolved packet core (vEPC) have become clear, and mobile networks operators are pretty much sold on the technology in theory.
As the technology side has been figured out and operators begin to plan commercial deployments of NFV and vEPC, however, discussion is starting to move toward operational requirements and challenges. Mobile network operators need to figure out how best to manage these new virtual network functions (VNFs) and the NFV infrastructure, and also how to modify the existing network operations model when these VNFs are deployed.
“These are understandable concerns since clearly there will be additional operational issues when this NFV-MANO [management and orchestration] network architecture is deployed,” noted Keith Allan, Director IP Mobile Core Product Strategy, Alcatel-Lucent, in a recent TechZine posting, vEPC: How to achieve operational elegance.
There are a number of new functional blocks and data repositories that come with this new model, including the MANO functions themselves, vEPC VNFs, element and network management systems (EMS/NMS), operational and business support systems (OSS/BSS), and NFV infrastructure.
For Allan, however, these concerns are real but solutions also exist for mobile network operators to deal with them.
Existing EMS/NMS can be combined together with an integrated NFV/SDN management solution and enable mobile operators to address NFV operational challenges while also being able to manage the existing purpose-built, product-based network using their current OSS/BSS, according to Allan.
This combined system, which Alcatel-Lucent business unit Nuage Networks and the Alcatel-Lucent CloudBand team are developing can deliver, enable workflow automation with push button VNF instantiation and elasticity, automates service chaining via SDN, and brings network function orchestration to coordinate multiple virtual and physical network functions.
How this is done is through dividing into three well-established management domains: virtual machine orchestration and VNF/VNFC life cycle management, network connectivity orchestration, and network function orchestration.
“This combined element and network management solution for NFV/SDN delivers the operational elegance that mobile operators need to reduce complexity,” noted Allan in his blog post, “and it opens the door for innovation to provide new services through automation.”
As operators move from testing to commercial rollout, such solutions will increasingly rise in importance.
]]>