Category Archives: Virtualisation

BCN cloud coverage moves to Telecoms.com

Erfolg Richtung PfeilAs the cloud and telecoms sectors move ever closer together, thanks to the growing influence of virtualisation in managing networks and the emergence of IoT, the decision has been made to consolidate the coverage of both under one brand – Telecoms.com.

It is already clear that the telecoms business for the foreseeable future will be dominated by three major themes: 5G, IoT and cloud. The eventual 5G standard will lean heavily on the cloud via technologies such as NFV and SDN, while IoT will be entirely dependent on the cloud to assimilate the massive amounts of data generated by billions of IoT devices and then turn it into useful actions.

Business Cloud News has been excellently led by first Jonathan Brandon and more recently Jamie Davies. Jamie is continuing his great work on Telecoms.com as Deputy Editor and cloud specialist. On top of that we will be working closely with our sister title Light Reading, which has also established cloud and virtualisation as a cornerstone  of its coverage.

We would like to thank BCN’s thousands of readers for their loyalty and support over the past few years and hope you continue to enjoy the cloud coverage on Telecoms.com. BCN newsletter subscribers will be transferred to the Telecoms.com mailing list but you are, of course, free to unsubscribe at any time, much as we hope you don’t.

The whole ICT sector is on the cusp of a uniquely exciting era and we aim to ensure Telecoms.com will remain your one-stop-shop for all the latest developments.

Three major run-time performance hurdles to avoid

Network Function VirtualisationEvery investment in major IT transformation arrives with its own set of structural and institutional challenges, not least migrating to a virtualized infrastructure. Get it wrong from the outset and you will be faced with the mother of all headaches. If the issue of performance is not addressed at inception, your data centre and its services – although being able to run – will be increasingly plagued by service problems, poor user satisfaction and inadequate Return on Investment. So, companies looking to transform their virtualised infrastructure should look before they leap and consider three major hurdles.

#1: The heavy lifting of remedial actions

Shift happens. Seasonal traffic spikes, buggy drivers, fluctuating head counts, updated applications, worn out drives and a thousand other factors all contribute to the constantly shifting foundation upon which your infrastructure must stand and thrive. But in any dynamic environment, the unforeseen throttle will inevitably occur and issues will need fixing.

End-user issue remediation actions are exceedingly important, so when a VDI (Virtual Desktop Interface) user calls or files a trouble ticket, IT must have the tools for translating complaints into troubleshooting. Pinpointing problem sources manually can take hours; sometimes even days or weeks, because virtual infrastructure tools tend to be designed for high-level observation, not granular examination of what exactly is, or isn’t, going on. If multiple problems crop up from different sources, the burden of remediation magnifies, potentially risking an organisation’s business continuity.

#2: Isolated end-user experience metrics

Amazon famously publicised how each 1/10th of a second of added site latency sacrificed 1% in sales. Nothing will persuade a worker not to use a cloud-based software tool, particularly on a new rollout, quite like staring through a login hourglass. Of course, management feels the same pain. Aberdeen Research found that poor application performance can slice up to 9% from corporate revenue.

Citrix proposes that there are five key, measurable metrics that constitute the lion’s share of poor user experiences within virtualised environments:

  • Launch time
  • Logon time
  • Load time
  • Latency
  • Operations such as printing, screen updates, etc.

Clearly, most issues will stem from obstructions in resource flow, but determining the root cause or causes of any decline in these metrics can be arduous.

#3: Inefficient allocation of IT resources

Virtual infrastructure resource utilisation wastes mountains of money when it runs too cold and throws up bottlenecks and failures when it runs too hot. Efficiency is the relationship between performance and resource utilisation and maximum efficiency does not mean maximum utilisation. The physical underpinnings of your virtual infrastructure are a mesh of CPU, memory, storage, network and – particularly in VDI and HPC deployments – GPU (Graphics Processing Unit) assets. The key is to deduce not only utilisation levels but overall efficiency within each resource type, network region and entire organisation.

The solution for companies looking to jump these hurdles is to invest in better predictive analytics-driven strategic levers. They enable the reduction or even elimination of persistent IT problems. To prevent the heavy lifting of remedial actions the tool needs to increase real-time prevention of problems and in cases when prevention is not enough, step in with automated remediation. Either way, the end result is massive savings in resolution time as well as a much improved chance for IT, and thus the whole organisation, to meet its SLAs (Service-Level Agreements). The ability to help pinpoint the problem’s source and whenever possible, cross-silo insights for balancing load optimisation to alleviate bottlenecks is crucial to prevent isolated end-user experience metrics.

In order to combat inefficient allocation of IT resources you must be able to monitor the real-time characteristics of these physical resources and how end-points and applications are using them across the network. Clarity of vision is a strategic lever not only to utilisation levels but overall efficiency within each resource type, network region and entire organisation.

Written by Atchison Frazer, CMO at Xangati

Cisco launches Digital Network Architecture virtualization platform

Network Function VirtualisationCisco has launched a new system which aims to virtualize every conceivable network function possible for clients and take them through the painful process of digital transformation, reports Telecoms.com.

The networking vendor has announced its new Digital Network Architecture (DNA), which it describes as an open, software driven framework. The DNA will complement and extend the policies of its datacentre based Application Centric Infrastructure (ACI) technology throughout the entire network, Cisco says.

Whereas ACI software defined the network, DNA will help enterprises to define everything from the campus to the branch, whether the network is wired or wireless, at the core or at the edge, says Cisco. DNA will sit within the Cisco ONE Software family, in order to simplify software licensing and help protect investments by providing continuity.

The Cisco DNA is built on five guiding principles, which can be summarized as virtualize everything, automate management, analyse everything everywhere, one policy for the entire network and keep every layer of networking as open and extensible as possible.

The mission to virtualising everything that can be possibly software defined will maximise the options for telcos and all enterprises. This gives the clients the choice to run any service anywhere, independent of the underlying platform, be it physical or virtual, on premise or in the cloud, says Cisco. Yesterday Cisco announced the acquisition of Leaba Semiconductor, which specialises in networking semi-conductors which could play a central role in the virtualisation of networking functions and maximise the possibilities for embedding virtualised functions.

Automating network management will maximise the speed and efficiency of the virtualised functions of an enterprise, but this may be regulated by the third important DNA principle, the need to have pervasive analytics. Analytics will provide the checks and balances needed to keep the network and IT infrastructure meeting its performance potential. Similarly, a virtualised network can only be an efficient cloud if service management from the cloud can unify policy and orchestration across the network. On Monday BCN reported how Cisco plans to buy cloud orchestration specialist CliQr.

The key to preventing network sclerosis is keeping everything open and accessible, which is why Cisco’s fifth guiding principle for DNA is to keep everything open, extensible and programmable at every layer, so that Cisco and third party technology can be integrated.

OPNFV announces second major release – Brahmaputra

Digital illustration of Cloud computing devicesThe Linux Foundation-inspired OPNFV Project has taken a new step closer to its ideal of network liberalisation with a new release of its software.

Network Function Virtualisation (NFV), the telecoms industry’s answer to the Stock Market’s Big Bang, aims to open the market for creating software that runs the multitude of functions within any network. The OPNFV Project aims to create a carrier-grade, integrated, open source platform that uses NFV to create telecoms networks that are infinitely more flexible and adaptable than the traditional proprietary systems that locked the software within the rigid backbone of telecoms hardware.

The Project has announced the availability of new improved version of its original offering, code-named Arno, which Telecoms.com reported on in June 2015. The new release, Brahmaputra, offers a more comprehensive standard of tools for testing NFV functionality and use cases. Brahmaputra is OPNFV’s first full experience with a massively parallel simultaneous release process and helps developers to collaborate with upstream communities. By encouraging group collaboration on feature development and addressing multiple technology components across the ecosystem, the Project aims to improve the stability, performance and automation of the system, and to consolidate its features.

The extent of collaboration is ambitious, since OPNFV aims to bring together at least 165 developers from network operators, solution providers and vendors. The focus of their joint efforts will be on integration, deployment and the testing of upstream components to meet NFV’s needs. During the integration process to create the Brahmaputra release, code was contributed by programme writers in the OpenStack, OpenDaylight, OpenContrail ONOS and ETSI developer communities. Meanwhile, there were 30 different projects accepted which created new powers, specifications and community resources to the system.

Among the improvements are Layer 3 VPN instantiation and configuration, initial support for IPv6 deployment and testing in IPv6 environments, better fault detection and recovery, performance boosts through data plane acceleration and much fuller infrastructure testing.

“The strength of any open source project depends on the community developing it,” said OPNFV director Heather Kirksey, “with an entire industry involved in the development of NFV, we’re seeing more collaboration and the strides we made in Brahmaputra create a framework for even more developers to come together.”

HPE scoops two telco client wins for cloud service projects

HPE office logoHewlett Packard Enterprise (HPE) has announced partnerships with telcos Swisscom and Telecom Italia subsidiary Telecom Personal to share its cloud service expertise and boost its presence in the comms industry.

In the Swisscom project HPE’s brief is to impose a network function virtualization (NFV) discipline on the IT and telecoms infrastructure, using its OpenNFV systems. Swisscom claims it is one of the world’s first communication service providers (CSPs) to pioneer the use of NFV to offer virtual customer premise equipment (vCPE) to its business customers.

In January BCN reported that HPE has launched an initiative to simplify hybrid cloud management for telcos using a new Service Director offering. Among the productivity benefits mooted for HPE Service Director 1.0 was options for pre-configured systems to address specific use cases as extensions to the base product, starting with HPE Service Director for vCPE 1.0.

In the Swisscom project HPE will use its HPE Virtual Services Router and HPE Technology Services in tandem with Service Director to create Swisscom’s new vCPE model. The objective is to allow Swisscom to manage its customers’ network infrastructure from a centralised location and provide networking services on-demand. This will cut costs for the telco, speed up service provision and boost the availability of services. It could also, claims CPE, make it easier to create new services in future.

Argentina based Telecom Personal has asked HPE to modernise its network in order to use 4G/LTE technologies to cater for an increasing appetite for data services among subscribers. HPE has been appointed to re-engineer the infrastructure and expand and upgrade part its network core. The success of the project will be judged on whether HPE can give a measurable improvement in service experience, network speeds and capacity, according to Paolo Perfetti, Telecom Personal’s CTO.

Yesterday BCN reported that HPE has launched AppPulse Trace, a service that developer clients can use to monitor their cloud app performances.

Wind River launches vCPE NFV platform

Network Function VirtualisationIntel subsidiary Wind River has added to its NFV portfolio with the announcement of a new product which focuses on the virtualization of customer premises equipment (vCPE). The vendor claims it has strengthened its VNF offering to bring initiatives like vCPE management closer to reality.

Wind River’s new Titanium Server CPE software adds to the vendor’s existing NFV portfolio. vCPE is one of the primary NFV use cases mooted by the telecoms industry; the improvements to Wind River’s software could mean all functions traditionally configured on each piece of hardware at the customer’s premises can now be handled centrally and instantly replicated across the cloud.

According to Wind River, the improvements to the Titanium Server portfolio (Titanium Server and Titanium Server CPE) will create faster network performance, allow networks to ramp up in size instantly and simplify the commissioning of new additions to any network. The vendor also claims the Titanium Server portfolio’s new software stands up to the carrier grade needed for telecom networks.

The cost of administering moves and changes, which is often time consuming and labour intensive, is one of the areas that telcos view as ripe for efficiency improvements. However, the virtualization of CPE functions has proved a challenging ambition to fulfil. The replacement of a system which individually configures each physical appliance (such as a switch or handset) with a centrally managed one could slash operating expenditure (OPEX). The savings would come by substituting the work of multiple physical configurations, requiring a site visit, with a single software change that can be published to all devices on all points of the telecoms cloud, using the new Titanium Server CPE platform, according to Wind River.

Among the new features added to Titanium Server are dynamic CPU scaling, greater system scalability, a virtual switch packet trace tool and the capacity for bulk provisioning and automated deployment. The new release also now supports QinQ tunnelling, IPv6 support for all interfaces and has updated support for all the latest high-performance network interface cards (NICs). It also supports VNF access to hardware acceleration devices, such as the Intel Communications Chipset 8925 to 8955 Series.

The new incarnation of Titanium Server CPE launch follows collaboration with NFV software partners Brocade, Check Point, InfoVista and Riverbed.

Titanium Server and Titanium Server CPE will be showcased at Mobile World Congress 2016.

Openstack targets telcos with NFV push

Digital illustration of Cloud computing devicesA new report indicates that there could be a boom in network function virtualisation projects this year, with NFV the second most popular subject of research after containers, reports Telecoms.com.

According to a report from the OpenStack Foundation, only container technology is under closer scrutiny than NFV by technology buyers and decision makers in the world’s enterprises and service providers.

The paper, Accelerating NFV Delivery with OpenStack, reports on the findings of the foundation’s most recent user survey, in which 76 per cent of those questioned identified an important telecoms function that had to be addressed through virtualisation. Of the OpenStack user base 12% were traditional telcos and another 64% were companies that now include telecoms as part of their roster of services, such as the categories of cable TV and ISP companies, telco and networking and data centre/co-location companies.

By comparison, an OpenStack user survey in 2014 suggested its user base of telcos was much smaller, the Foundation says, and only an elite of global telcos, such as NTT and Deutsche Telekom, were investigating NFV use. Since then there has been a surge in interest, it reports, with

increasing numbers of telecom-specific NFV features, such as support for multiple IPv6 prefixes, being requested or submitted by OpenStack users.

Container technology information is even more sought after than NFV, according to OpenStack, but the two issues are not mutually exclusive. Sources have speculated that the technologies may be used in tandem as OpenStack is the foundation of rationalising the hybrid nature of most telco’s infrastructure.

According to the paper’s executive summary OpenStack could provide cost effective route to the creation of private clouds without vendor lock-in, since proprietary hardware is becoming associated with NFV.

“While the interoperability between NFV infrastructure platforms that use OpenStack is still a work in progress, the majority of configurations surpass expectations,” concluded the paper co-authored by Kathy Cacciatore, the OpenStack Foundation’s Consulting Marketing Manager.

Big Switch Networks wins $48.5M to bring SDN to telcos, data centres and enterprises

Network Function VirtualisationSanta Clara based software defined networking vendor Big Switch Networks (BSN) has won another $48.5 million to bring its bare metal networking fabrics to new markets, reports Telecoms.com.

The networking specialist, which has now received $93.5m since its launch in 2010, aims to use the new cash injection to fund more R&D and to create news sales and marketing channels in the Europe, Asia Pacific, the UAE and the US.

Investors from Morgenthaler Ventures, Silver Lake Waterman, Index Ventures, Khosla Ventures, Redpoint Ventures, Accton, CID Group and MSD Capital put the cash up after hearing how the company achieved 300% growth last year. Its two technology inventions have found three popular use cases among telecoms carriers, data centre companies, service providers and enterprises.

BSN offers clients a Big Monitoring Fabric and a Big Cloud Fabric, both of which are based on bare metal software defined networking principles, with a centralised product-specific controller managing a network of bare-metal Ethernet switches. The controller, the managed switches, and the links connecting to them form the network fabric. BSN defines the software for a centralised controller running on industry standard servers and the operating system that runs on the bare-metal Ethernet switches. The Big Monitoring Fabric, which connects networks with monitoring tools and Big Cloud Fabric, which provides software defined management of switching fabrics in data centres, have won telco and data centre clients in the US, APAC and EMEA. Its main vertical markets are telcos and IT, financial services, government, service providers and higher education.

In addition to the extra funding, BSN announced that it has recruited former NetApp CEO Dan Warmenhoven and venture capitalist Gary Morgenthaler, who have both steered companies through the transition that comes with rapid expansion.

According to IHS Research, the percentage of users of software defined networking in enterprise communications will grow from 6% to 23% in 2016. It also estimates that spending on data centre networking will reach $13 billion in 2019, up from $781 million in 2014.

“Nobody can ignore the advantages of software defined networking,” said Shawn O’Neill, MD of one Big Switch’s venture partners Silver Lake Waterman.

Big Switch is fundamentally changing the economics of data centre networking and SDN, claimed  another investor, Mike Volpi, a partner at Index Ventures. “This financing will fuel significant go-to-market acceleration and geographic expansion,” said Volpi.

New Service Director from HPE could simplify hybrid cloud management for telcos

HPE street logoHPE claims its new Service Director system could put comms service providers back in control of their increasingly complex hybrid computing estates. It aims to achieve this by simplifying the management of network function virtualisation (NFV).

HPE claims that Service Director will automate many of the new management tasks that have been created by the expanding cloud environment and provide a simpler system of navigation for all the different functions that have to be monitored and managed. The new offering builds on HPE NFV Director’s management and orchestration (MANO) capacity and bridges existing physical and new virtualized environments.

As virtualisation has expanded it has extended beyond the remit of current generations of operations support systems (OSS) and the coexistence of physical and virtual infrastructure can introduce obstacles that slow the CSPs down, HPE said. It claims the Service Director will help CSPs roll out new offerings quicker.

The main benefits of the system outlined by HPE are automation of operations, shared information, flexible modelling of services and openness. With a single view of the entire infrastructure and dynamic service descriptors, it aims to make it easier to spot problems and create new services, HPE claims. As an open system the Service Director platform will have interfaces to any new third party software defined networking controllers and policy engines.

Since there is no such thing as a green field NFV set up there has to be a system to rationalise the legacy systems and the new virtualised estate, said David Sliter, HPE’s comms VP. “Service Director is a transformational change in the relationship between assurance and fulfilment, allowing the OSS resource pool to be treated, automated and managed as a service,” said Sliter.

The telecoms industry needs an omnipotent service orchestration system that can span every existing NFV MANO and OSS silo, according to analyst Caroline Chappell, principal analyst of NFV and Cloud for Heavy Reading. A model-driven, fulfilment and assurance system like Service Director could speed up the delivery of services across a hybrid physical and virtual network, Chappell said.

HPE Service Director 1.0 will be available worldwide in early 2016, with options for pre-configured systems to address specific use cases as extensions to the base product, starting with HPE Service Director for vCPE 1.0.

Cisco boosts SDN range with ACI update

Cisco corporateCisco claims that customers can take a further step towards network automation as it launched a new release of Application Centric Infrastructure (ACI) software to its software defined networking range.

Despite massive demand there are only 5% of networks being automated, according to Cisco’s own customer feedback. In response it has moved to simplify the task by making it easier to address all the various autonomous segments of any complicated network infrastructure.

The new software revision of ACI makes it capable of microsegmentation of both physical (i.e. bare metal) applications and virtualized applications, which are separated from the hardware by virtual operating systems such as VMware VDS and Microsoft Hyper-V. By extending ACI across multi-site environments it will enable cloud operators and network managers to devise policy-driven automation of multiple data centres.

In addition, Cisco claimed it has paved the way for integration with Docker containers through its contributions to open source. This, it said, means customers can get a consistent policy model and have more options to choose from when using the Cisco Application Policy Infrastructure Controller (APIC).

ACI now supports automated service insertion for any third party service running between layers four and seven on the network stack, it said. More support will be put behind cloud automation tools like VMware vRealize Automation and OpenStack, including open standards-based Opflex support with Open vSwitch (OVS).

The ACI ecosystem now makes the automation of entire application suites possible, including Platform as a Service (PAAS) and Software as a Service (SAAS) and there are now over 5000 Nexus 9000 ACI-ready customers using Cisco’s open platform it said.

“Customers tell me that only five to ten percent of their networks are automated today,” said Soni Jiandani, SVP at Cisco. Though they are eager to adopt comprehensive automation for their networks and network services through a single pane of management, they haven’t managed it yet. However, since several ACI customers have achieved full this could be the next step, said Jiandani.