Friday, July 14, 2017

Cisco – fog computing power combined with IoT management

by James E. Carroll

This week brings the 28th occurrence of Cisco Live!, the company's big technology and customer showcase (previously known as Networkers) that brings thousands of attendees to the Mandalay Bay Convention Center in Las Vegas. The big news was unveiled at a press conference last week in San Francisco, with Cisco seeking to 'reinvent networking' by applying machine learning to new network platforms powered by custom ASICs and secured via the Cisco Talos threat management platform. With this big news already out the door, the Cisco Live! event can focus on other topics including the company’s budding partnerships with Apple and Ericsson, as well as the theme of edge computing, which is gaining momentum across the industry.

Bringing compute power to the edge

As the number of connected devices grows, the flow of data from the edge of the network to the core increases. In his Cisco Live! key note, company CEO Chuck Robbins observed that we are already in a multi-cloud world. In the IoT context, this is good because data will flow from the edge to multiple cores, thus avoiding the problem of single point for concentration. Another increasingly popular term for this is fog computing. The recently published Cisco Visual Networking Index (VNI) predicts that M2M connections will represent 46% of connected devices by 2020.

There is the possibility to distribute general purpose computing resources at the edge of the network. By processing data at the edge, we can determine which data needs to be transported to a core data centre and which data can be acted on locally. Once it has been determined which data should be sent to the core, there is a need to transport it securely. Devices need to be managed and traffic analysed to understand the impact on network performance. Cisco already provides these capabilities through its edge devices and with its Cisco Jasper platform. The idea is to expand the Jasper franchise, thus building a business with predictable and recurring revenue in a market segment with extremely fast growth.

The Jasper connection and Edge processing

In February 2016, Cisco acquired Jasper Technologies for $1.4 billion in cash and assumed equity awards. The deal was finalised the following month. The Santa Clara, California-based start-up, headed by Jahangir Mohammed, a serial entrepreneur known for his previous company Kineto Wireless, developed a cloud-based Software as a Service (SaaS) platform to manage connected devices. At the time the Cisco acquisition was announced, Jasper claimed to have 3,500 enterprise customers and 27 service provider customers. Jasper had been working with AT&T on connected device management since at least 2009.

Currently, Cisco Jasper claims over 11,000 enterprises and 43 million devices using its IoT management platform. One notable customer is Amazon which manages Kindle devices using Cisco Jasper. Listed service provider customers include AT&T, Telefonica, Singtel, KT, China Unicom, NTT Docomo, VimpleCom, Vivo, Bell, Telus, Rogers, Comcast, Optus, Telstra, Etisalat, TIM, O2, and many others. In the automotive sector, Cisco Jasper is used by a number of manufacturers, including Ford, GM, Subaru, and VW. Several of these players, notably GM, are now including connected car service as a basic feature for many models, which means tens of thousands of new connections every month for a single customer. At Cisco Live!, Chuck Robbins said the 1.7 million new devices are being managed by Cisco Jasper every month. In fact, Cisco now claims this to be the world’s largest IoT connectivity management platform.

The idea of actionable intelligence at the network edge is a powerful one. One example cited at Cisco Live is Chevron, the multinational energy company based in San Ramon, California. As one of the largest oil companies worldwide, Chevron has drilling operations in locations across the west coast of North America, the U.S. gulf coast, Australia, Nigeria, Angola, Kazakhstan, and other locations. In the U.S., Chevron operates approximately 11,000 oil and natural gas wells. Its $43 billion Gorgon Gas Project in rural Western Australia is the largest liquified natural gas (LNG) initiative in the world. Drilling is extremely data intensive activity, and rigs and wells are being equipped with every possible type of automated sensors, resulting in a tremendous flow of data. Previously, achieving actionable intelligence from a rig out in the field typically took 2 weeks. With its IoT edge processing, Cisco said it is able cut actionable intelligence response time to under 2 seconds.

Chevron is an interesting example of edge computing which could be quite powerful in many vertical applications. For service providers, although edge processing might reduce the total volume of data traversing the network, it makes the flow of sensor data more predictable. It also makes the network more intelligent and therefore of higher value to the end customer. Service providers working on IoT will probably look to follow this model. Cisco says that by transforming IoT sensor data, fog applications like this can also benefit its ecosystem partners. The distributed IoT network can function as the compute environment for fog applications. Other examples of deployed fog applications include site asset management, energy monitoring, and smart parking.

The first evolution of Cisco Jasper

Fifteen months after completing the Jasper acquisition, Cisco is announcing the first upgrade to the platform under its ownership. Cisco Jasper Control Center 7.0 focuses on lifecycle automation, analytics, security, cost management and back-end integration. Enhancements in this release include better tools for traffic segmentation to allow deeper revenue and service models, as well tools to prevent attacks at the DNS layer. It also offers:

·         Advanced capabilities via a new tier of the platform – Control Center Advanced – which caters to customers with more sophisticated deployments that need greater capabilities, including advanced security solutions, automation and analytics.

·         Premium services, a threat protection and smart security (TPSS) service that provides an IoT-specific solution to protect against malware and other cybersecurity threats built on Cisco Umbrella; traffic segmentation provides a new solution to support different types of revenue generation models, with Cisco Jasper service providers can customize premium services to meet their customers’ specific needs.

·         LPWAN Support: Control Center 7. extends the platform's capabilities, including the platform's global reach and scale via support for cellular networks to low power devices via support for multiple LPWAN technologies including NB-IoT and LTE-M.

Swisscom to trial NB-IoT and Cat-M1 as part of 5G development

Swisscom has detailed efforts in the development of 5G technology, specifically stating that it plans to launch two new access technologies for the Internet of Things (IoT) next year: Narrow Band-IoT and LTE Cat-M1.

The access technologies are based on mobile supplements to the existing low power network (LPN), and Swisscom claims it will be the first provider in Switzerland to conduct field trials of network slicing in 2018. This technology is designed to enable individual 5G and 4G applications to be allocated guaranteed network resources in future.

Swisscom is also testing 5G speeds in a test environment in Zurich. Based on technology from Ericsson, the trials have achieved transmission capacity of over 20 Gbit/s capacity in a radio cell.

Swisscom noted that IoT is a key element in digitisation and will result in increasing mobile data volumes, therefore it is implementing a national expansion for IoT. In 2017, the low power network (LPN) will be made available to 90% of the Swiss population. Swisscom is also expanding its IoT portfolio via two access technologies based on 3GPP mobile communications - Narrow Band-IoT (NB-IoT) and LTE Cat-M1.

Swisscom plans to begin initial tests of NB-IoT and LTE Cat-M1 with pilot customers before the end of 2017, with a commercial launch scheduled for 2018.

To leverage the benefits of 5G, Swisscom is also implementing network function virtualisation (NFV) technology. In May, Swisscom launched an NFV service for companies. NFV allows the network load to be configured to ensure that applications have the required resources. To achieve this, network slicing ensures applications used in industrial communications are guaranteed network resources. Swisscom will test a prototype with partner Ypsomed and carry out the first field trials in 2018.

As part of 5G development, Swisscom is trialling a mobile base station with antenna and two terminal devices at Ericsson’s 5G development centre in Sweden. The 5G prototype enables speeds of up to 10 Gbit/s to be achieved in parallel on the mobile network.

Swisscom is also expanding its 4G network and will phase out its 2G network by the end of 2020. Swisscom currently offers 4G+ to 40% of the Swiss population with speeds of up to 300 Mbit/s. This is due to rise to 67% by the end of 2017. Swisscom also covers 15% of the population with speeds of up to 450 Mbit/s.

Vitalpointz Targets the Edge

We have seen the tremendous potential for innovation, cost saving, and flexibility being unleashed by the public clouds.  The hyperscale data centers of the top three public cloud providers are marvels to behold. Private cloud data centers and hybrid cloud architectures are also on the rise as Fortune 500 companies shift their IT spending to take advantage of this trend.

We are also now witnessing the long-predicted rise in the Internet of Things.  IDC’s recent Worldwide Semiannual Internet of Things Spending Guide predicts spending on IoT will reach $800 billion this year, up 16.7% year over year, and rising to nearly $1.4 trillion in 2021.  IDC breaks down 2017 investments in IoT as follows: manufacturing operations ($105 billion), freight monitoring ($50 billion), and production asset management ($45 billion), smart grid technologies for electricity, gas and water and smart building technologies ($56 billion and $40 billion, respectively). Looking to 2021, IDC expects these use cases will remain the largest areas of IoT spending. The use cases that will see the fastest spending growth are airport facilities automation (33.4% CAGR), electric vehicle charging (21.1% CAGR), and in-store contextual marketing (20.2% CAGR).


The Need for Edge Computing

At the intersection of these two trends is a new opportunity that is just beginning to catch the interest of Silicon Valley – edge computing, sometimes also called fog computing. The primary idea here is that Industrial Internet of Things (IIoT) applications will benefit from both cloud infrastructure and local compute/storage resources. Centralized controllers in the cloud could be used for provisioning, performance monitoring, billing, and big data analytics. Real-time control of the application and its associated physical devices would be retained by an “edge” processing/storage unit.

This will drive the development of small server farms, or “cloudlets”, located in-building, on-campus, or in a metro area data center.  Google recently disclosed plans for more data centers in city centers rather than solely hyperscale campuses in remote locations. AWS is promoting its “Greengrass” project, software for running local compute, messaging, data caching, and sync capabilities for connected device. Greengrass runs locally and the AWS cloud provides management, analytics, and durable storage.

The communication service providers have their own variation for this general concept - Central Office Re-architected as a Data Center (CORD).  Under the Linux Foundation, CORD is now an independent open source project aimed at leveraging the elasticity of commodity clouds and merchant silicon for a new generation of smaller and more efficient central offices. Backers include Google, Radisys, Samsung Electronics, AT&T, China Unicom, Google, NTT Communications, SK Telecom, and Verizon, vendors Ciena, Cisco, Fujitsu, Intel, NEC, Nokia, etc.

The Vitalpointz Application Forking Engine

Vitalpointz (vitalpointz.net) is a Silicon Valley-based start-up with R&D operations in Bangalore, that has just announced its entrance into this market. The company is led by veteran successful entrepreneur Ravi Medikonda, who previously headed Vistapointe, a start-up that specialized in cloud-based and real-time network monitoring solutions for mobile operators. Vistapointe developed data extraction, analysis and insight generation technologies that enabled mobile operators to gain visibility into their mobile networks. The solution leveraged Network Functions Virtualization (NFV) architecture, enabling it to run in a telco cloud.  Brocade acquired Vistapointe in 2014. The Vistapointe team went on to become Brocade’s Network Visibility and Analytics business unit, establishing accounts with major North American and Japanese mobile operators.  With Broadcom’s $5.9 billion acquisition of Brocade and subsequent divestitures of many business units, the time seemed right to pursue the new edge opportunity.

“We see a distinct opportunity for a better edge computing paradigm,” says Vitalpointz’ Ravi Medikonda. “Application developers really should not have to know specifically what hardware resources are available locally versus in the cloud.  Our forking engine will automatically direct traffic to where it can be best processed. In many cases, that might be a nearby CORD or on-prem server, but it might be the public or private cloud.”

Applications are driven by multiple functional modules, also known as micro-services, which can exist in different locations (VMs, container, across racks, across data centers, etc.). We also know that application deployment has changed to a SAAS multi-tenant model.  The same deployment of "Office 365" can serve multiple companies and customers.  So, the ability to specifically manage an application by host or an IP-address is not possible.

The patent pending Vitalpointz Application Fork Engine (VAFE) technology will enable applications to run “as is” across the cloud and cloudlet without any configuration change. The company says its VAFE technology will benefit several use cases that require quick responsiveness, low latency and near real-time operation. VAFE can be embedded in x.86 platforms, VMs, processor boards in Layer-2 DC switches or IIOT gateways.

Examples could include context-aware services and location-aware services, asset tracking, video surveillance, connected cars, augmented and virtual reality, etc.  Think of a hotel that is installing NFC-enabled door locks on its customer rooms.  When a new reservation is booked online, a room suite is automatically assigned and a unique room access code is generated and emailed to the guest. This part of the booking is handled by the hotel management application in the cloud. When the guest arrives at the hotel on the day of the booking, he or she may proceed directly to the reserved room, which opens when their NFC-enabled phone is touched to the door lock. The authenticated door opening transaction is processed locally rather than in the cloud data center which could be thousands of miles away.


The Vitalpointz founding team has played the Silicon Valley & Bangalore start-up game before with a successful outcome. A promising market opportunity has been identified and key intellectual property is under development. As is often the case, it is the focused engineering teams who have worked together in the past that gain a first-to-market advantage over the large vendors.

See also