Mid-band spectrum shortages in the US was the main thrust of the 5G Americas sponsors. The idea is that the rest of the world has lots of mid-band spectrum available and service providers in countries that could be considered economic leaders (Japan, Korea, China, Western European countries) have plenty of available mid-band spectrum that is ideal for 5G, while the US does not.  This group at 5G Americas, which includes service providers, vendors and standards bodies, is saying that US leadership in cellular infrastructure and the entire app economy that relies upon it may be at risk as 5G get deployed.

Other topics discussed:  AT&T is currently out for bid on its 5GC infrastructure, and this caused some interesting posturing by the vendor attendees (like Ericsson, Nokia, Cisco, Mavenir) at this conference, with each trying to identifying their strengths.  It seems the consensus is that all mobile operators in the US market are using Option 3X, an EPC anchoring system.  And, the consensus seems to be that US operators will need to move to 5GC once most traffic is coming over 5G Radio (“New Radio”).  Vendor selection appears an open field, once again, as 5GC has 13 different microservices, each which could theoretically be parsed out to different vendors.  Operators are saying, though, that while this multi-vendor selection may lead to savings on purchasing, it will increase integration spending, so these two have to be balanced out. 

Mobile Edge Computing:  The consensus is that a 50 mile radius (or others are saying 100 km) is considered the ‘edge,’ or the ‘low latency’ zone.  We expect, however, that the data forwarding plane (‘user plane’) will be distributed to, say, 100 locations within a territory like the US market, while the control plane will be much more centralized (perhaps as centralized as it is currently, where it might be considered to be like 1/4th the number of locations). 
​
CBRS.  The consensus is that testing will be done by mid November 2018 and Initial Commercial Deployment by 1Q19, potentially spilling into 2Q19.  PAL auctions are expected by attendees to be a 2019 event, with 2020 traffic running on PAL spectrum. Commscope represented the views from a SAS standpoint for this discussion.There were discussions about the C-Band (6 Ghz) potentially using the same type of Automatic Frequency Coordination system, but the consensus is that it is too early to declare that the path forward.

We attended Mobile World Congress Americas (MWCa) in Los Angeles, CA this week, as well as the AT&T Spark event in San Francisco.  Since 5G is launching first the US, these two events became the public events where significant 5G-related announcements happened.

• Verizon.  Will launch 5G Fixed Wireless Access (FWA) on October 1 in four markets: Houston, Indianapolis, Los Angeles and Sacramento.
• AT&T.  The company reiterated its own 5G plans (mobile 5G by year-end 2018 in cities such as Atlanta, Charlotte, Dallas, Indianapolis, Oklahoma City, Raleigh, Waco, Houston, Jacksonville, Lousville, New Orleans and San Antonia), plus it made some announcements like that it is beginning 5G-ready CBRS equipment testing (using Samsung CBRS equipment and CommScope as SAS provider).  Also, at the Spark event on Monday, the company announced three strategic telecom equipment suppliers, Ericsson, Nokia and Samsung. 
• T-Mobile.  Announced that it had completed a Cisco vEPC system (upgradeable to 5G Core) carrying traffic for 70M users that was from Cisco.  It also announced that it signed a $3.5B 5G agreement with Ericsson.  This is in addition to the July 30 announcement made earlier with Nokia for $3.5B, as well.  Generally, the company has set expectations as recently as September 10 that it will provide nationwide 5G by the year 2020.
• Sprint.  Announced that it had demonstrated a 5G NR connection of Massive MIMO with Nokia equipment.

Additionally, discussions about spectrum in the US market were very active discussions.  Some points we picked up on:

• No new mid-band auctions will occur in the US market for another 2-3 years, so this means that new capacity is going to come from LAA (just announced on the iPhone Xs this week, as well) and from CBRS (discussed above).
• The "who has the fastest 5G throughput" battle will be won at the millimeter wave.  In other words, using millimeter was, speeds as high as 10 Gbps are possible, but with mid-band (1-6 Ghz), where LTE is currently deployed, cannot go much over 2 Gbps.  So, to beat the Ookla Speed Test, the mobile operators who deploy mmWave early will get a leg up.  However, in order to deploy mmWave, these have to be small-cells that are within 100 meters of users.  Since it is so difficult to get real-estate rights and backhaul for small cells, this is going to be a big challenge.  Nevertheless, this is how the battle will be won.
• T-Mobile's 600 Mhz rollout is now in 1,250 cities.  The company will eventually enable 600 Mhz 5G.  600 Mhz should dramatically improve T-Mobile coverage because it is low-band spectrum.

Yesterday, Ericsson and Juniper announced plans to partner to tackle 5G transport challenges together.  Additionally, Ericsson announced a partnership with optical transport company, ECI.  The companies said that Ericsson's Router 6000 product family will focus on fronthaul and backhaul and edge compute and will be complemented by Juniper MX, PTX and SRX Series portfolios providing edge, core and security capabilities.  Additionally, in today's presentation to analysts, Ericsson showed that the overall transport capability from Ericsson will also include Optical Transport from both ECI Telecom and Ciena.  The Ericsson Network Manager will be capable of managing not only Ericsson products, but also the three families of products from Juniper.

ECI will be used for Metro, and will be completely integrated with the Ericsson OSS platform.  The company conceded that it is de-emphasizing its in-house metro optical product line and focusing on ECI.  Ciena will be used for long-haul transport predominantly, mainly the Ciena 6500 product.  Ericsson concedes that both ECI and Ciena products can move to other domains.

To enable automation for 5G, Ericsson can only guarantee that networks working with Ericsson and its partners can successfully be automated.  This automation / partnership topic illustrates just how complex it will be to make 5G networks work properly.

Recall that Ericsson and Juniper had a close partnership that goes back many years.  In fact, the Juniper router predecessor to the MX was used as the underlying hardware for Ericsson's very important GGSN/SGSN and later EPC capabilities. Ericsson eventually replaced the Juniper product for EPC with its own Router 6000.  Subsequently, Ericsson announced a partnership with Cisco, which was more general in nature.  That relationship did not result in much tangible progress, from what we have learned; and the relationship was done under previous management teams for both Cisco and Ericsson.  Ericsson explained today in its analyst briefing that Cisco and Ericsson compete in several areas.  So, this new Juniper relationship is important in that it re-kindles an old relationship and plays to both companies' strengths.

We attended the Nokia Fixed Broadband group analyst meeting in Tokyo last week and found that the company’s portfolio is expanding rapidly.  The company’s view is that capital spending at its customers will be flat in future, and therefore the company is taking the approach to grow its portfolio beyond its core DSL and Optical core to potentially allow it grow revenues.  It has expanded its product line to include cable (from Gainspeed acquisition), more recently home WiFi (augmented through acquisitions) and is rolling out Fixed Wireless Access (FWA).  The company is also predicting that the Indian subcontinent and fixed network wholesaling will become significant opportunities for the Nokia Fixed Network group.

• Cable.  Admits that it took a year longer for DAA to develop.  In deployment phase at WOW! And we get the sense that there are some significant wins that Nokia could announce in the future.  It claims to be the only vendor with DAA available today.
• Fixed Wireless Access (FWA).  The company is offering two versions of FWA: The company is offering two versions of FWA: (a) 3GPP radio access network based and (d) unlicensed WiGig (60 Ghz WiFi-based).  The company expects 5G-based solutions to hit the market in 2019.  The company’s first FWA customer is an enterprise, using WiGig.  But, the main focus is on large Service Providers.  The company is not focusing on Wireless ISPs (WISPs).    
• WiFi.  Nokia is making a push to enhance its Broadband CPE portfolio to include more advanced WiFi capabilities.  It is launching a “mesh” capability that will become popular across its broadband CPE product line.  Mesh is a very popular router category that has been propelling growth for vendors such as eero and Google.

Additionally, Nokia emphasized an enabling technology to allow fixed network wholesaling - slicing.  The company has begun discussing Altiplano, an SDAN (Software Defined Access Network) system.  This allows multiple service providers to share the same fiber and equipment installed base.  The company’s view is that just like mobile network tower sharing happened to reduce capital spending, the same will happen with Fixed Broadband networks – but this requires a virtualization technology such as SDAN.

F5 kicked off its #F5Agility18 conference with over 1,300 attendees.  The company’s themes was “Any app, anywhere,” and words repeated included “Automation” and “as a service.”  The company is moving towards selling security as a standalone offering.  Accompanying the show, there was a press release announcing Gi LAN traffic management and Gi Firewall VNFs generally available September 2018 focused on mobile operators, available on capacity-based consumption model.  A striking difference between this year’s conference in Boston and last year’s conference in Chicago was that a significant portion of the management team is different: new CIO, CFO, regional VP, Chief Strategy Officer, two new SVP/GMs of business units, EVP HR, EVP Services.

Generally, the company is planning on delivering new capabilities:

• automation & orchestration,
• cloud-native applications service platforms (Packaged software and as a service for native cloud consumption experience (in next 12 months)
• new application security services.

Here are the three eras of market development for F5:

• 1996-2004.  Load balancer.
• 2005-2015- ADC market.
• 2016+    Multi-cloud application services

The company is putting significant investment into SP, software, automation.  “Any app, anywhere” vision.

​Growth strategy:

1. Drive share gain in physical ADC market.  F5 expects physical ADC down mid single digits Y/Y (5%).  Think ours will decline less fast b/c we have exposure to largest customers, have SP exposure who demand hardware (ViPRION) and because we innovate on hardware faster than competitors
2. Capturing growing demand for virtual ADCs and ADC-as-a-service.  Been growing vADC 20%/yr past few years; could accelerate.  Just now releasing purpose-built software (BigIP) and cloud-native software in early 2019. 
3. Expanding security with new multi-cloud solutions.  WAF enhancements. 
4. Creating long-term value through services. 
5. Driving margin expansion and earnings power.  Through scale.  40% software by 2022. 

Development timeline:

• F2018: Just announced orchestration / automation a few months ago for BigIP (BigIP Cloud edition).  Per app use-cases are enabled through this innovation.
• F2019: cloud-native app services.  Container-based, Kubernetes, cloud-native distributed ADC system.  Beta end this year, release 1H19.
• >F2019: ADC-as-a-service. 

Security strategy:

• F15-17: ADC-attached security
• F18-19: Standalone security offerings
• F20-22E: Multi-cloud application security

Fast-growing Ethernet Switch vendor Arista Networks announced plans to acquire Mojo Networks, an Enterprise-class WLAN vendor today.  The deal is expected to close in C3Q18.  This deal douses hopes that Arista may buy other WLAN vendors, in our view.  Mojo Networks is unique in the WLAN industry: (a) it has a different business model from competitors, and (b) as of Aug 2018, it manages more Enterprise-class Access Points using Artificial Intelligence than other vendors.  Arista says that it plans to use Mojo AI in its Cognitive Cloud Networking for Campus initiative.  

Mojo Networks has taken revenue for cloud-managed services and has not taken revenue from the sale of Access Points.  The Access Points are sold by distributors who make a small margin and drop ship them to Mojo Networks customers.  The Mojo Networks revenue from cloud-managed services is significant when compared to other vendors in the Enterprise-class WLAN market.  Additionally, because Mojo's Access Point selling partners charge only a small premium for the hardware, Mojo customers benefit with a lower total cost for the Access Points, than, say, Cisco or HPE Aruba Access Points.

​

Back in late January 2018, Aerohive announced its 802.11ax product line.  We have eagerly awaited for the company to ship and today, the company announced its first 802.11ax customer, Wellington College in the UK.  The company claims it is first to ship 802.11ax to a European customer; we are not aware of any others who have done so (but not every vendor puts a press release out about this kind of thing).  Still, quite an accomplishment.  We understand that the company plans to continue shipping more products in the coming weeks and that by "late summer" the 802.11ax product shipment status should be considered "shipping."  

We had a follow up call with the product managers at Aerohive about 802.11ax and got an update:

• Aerohive has had several beta customers already using its 802.11ax APs
• The company expects that the OFDMA features in 802.11ax will be more important than the MU-MIMO features in practical terms
• Target Wake Up Time and 2.4 Ghz support are important features supported in the first Wave of 802.11ax to support Internet of Things (IoT) devices
• Overlapping Basic Service Set (OBSS), 1024 QAM and 160 Mhz Channel width will all improve functionality of 802.11ax compared to its predecessors 802.11ac and 802.11n.

 

H3C and Qualcomm announced the planned availability of the WA6628 802.11ax Wireless LAN Access Point; it is slated for September 2018 shipments.  This new Access Point uses the Qualcomm IPQ8078 part number - we checked the Qualcomm website and we cannot find this part on the website.  So, we assume that this is a future product.  

While the H3C press release is not the first 802.11ax Access Point get announced (Huawei, for instance, announced its own .11ax product in early 2017), this is important because this press announcement has a specific shipment date - September 2018.

Additionally, it is generally expected that another set of announcements are imminent from vendors using Broadcom chips.  Looks like Qualcomm and H3C wanted to get ahead of those announcements.

Not quite a year ago, Cisco and Google announced a Cloud partnership. Today, at the very first keynote at Cisco Live 2018, Diane Greene, CEO of Google Cloud joined Chuck Robbins on stage to talk about the partnership, highlighting Kubernetes and a  unified security policy.   Both Chuck and Diane want a large ecosystem of partners and developers.  Later on, Chuck mentioned Cisco passing the 500K developer milestone for DEVNET.

Chuck touched a litter on routing, mentioning next-generation branch and highlighting intent based networking activity in the SP space.  For example, one of their SP customers updates 60,000 routers each night using automation. He then quickly got back to the Catalyst 9K switch, highlighted as the fastest ramping product ever in Cisco history.
 
Children’s Hospital of Los Angeles was highlighted as a customer of the Cat 9K. As a customer, they have over 35,000 connected devices.  They purchased over 2,000 WLAN APs and over 200 Cat 9300 switches. They are also deploying ISE and have done over 23,000 different device profiles/identifications and are track to start policy enforcement.  They are now in the process of deploying at their branch locations. They noted 550k blocked threats over the first few months of deployment.
 
My key takeaways are that there is an explosion of devices and data on the network, much of which is encrypted and a human can only do so much; thus the network must scale and automate.  Cisco is looking to use AI, automation, and its architecture to allow the customer to scale with those IoT devices and to have the network automate many tasks, especially around security. Monetization for Cisco will occur both in the hardware, but also in the solution sale.  An ideal customer would be end-to-end Cisco, but Cisco will also support open APIs in order to allow partners and customers to operate with their preferred solutions.

We attended the Ribbon Perspectives conference in Los Angeles this week.  Ribbon, the result of the merger between Sonus Networks and GENBAND, made the pitch that it is moving to the services oriented business of real-time communications over time instead of just being an infrastructure equipment company.  Ribbon has differentiated from competitors in a few different ways, including that it is the first vendor to bring GPU-based acceleration to the telecom network industry and that it is offering a white-label voice/Communications Service as a Platform (CPaaS) service.
​
Kevin Riley, CTO Ribbon commented that:

• Public cloud is viable for use in telecom
• The company has deployed Microservices SBC at edge and core in a large Tier 1 North American Service Provider

Disruptive technologies that Ribbon is pursuing:

• 5G
• IoT
• Big Data/ML
• uCPE
• SD-WAN
• GPU
• CPaaS

GPU helps transcoding, complex media in SBC environments.  Have invested for the past 2 years.  3.5x transcoding compared to DSP, 9x transcoding compared to CPU, 2x power consumption reduction.

Ribbon is using GPU in both private and public cloud, just won Japanese Interconnect at hyperscaler scaler for transcoding.  Will likely pursue GPU acceleration for cryptography, video and DPI.

Additionally, the company emphasized its relationship for its new services - a win with Softbank Japan chose Ribbon Protect.

Ribbon's bets:

• Analytics & Security
• Kandy.  Consumption-based.  Only "white label go to market strategy"
• GPU/NFV.  Only company in telecom with GPU expertise

David Walsh, Founder Kandy.
Customer wins discussed during presentation:

• Optus Loop
• IBM/Softlayer got Hertz as first customer, using Kandy.  IBM will become a major voice player, using Kandy, according to Ribbon
• City of Los Angeles.  Active Shooter application
• KPN CUBES
• IPC, 2,000 users now, rolling out further
• Etisalat

The company highlighted several features about Kandy:  IoS call-kit integration, Outlook integration, web browser integration, meeting collaboration, advanced conferencing, and an AI assistant capability.  The company reiterated its focus is on the enterprise market, unlike Twilio who it says it focused mainly on the consumer market.  The company declined to disclose revenues of its Kandy business.