Tag: edge computing

Articles

New Amazon Echo devices

Everything Amazon announced at its September event | Mashable

Amazon hardware event 2020: Everything the company announced today | Android Authority

Use of edge computing in new Echo devices

Amazon’s Alexa gets a new brain on Echo, becomes smarter via AI and aims for ambience | ZDNet

From the horse’s mouth

Amazon

Introducing the All-New Echo Family—Reimagined, Inside and Out (Press Release)

New Echo (Product Page with ordering opportunity)

New Echo Dot (Product Page with ordering opportunity)

New Echo Show 10 (Product Page with ordering opportunity)

My Comments

Amazon are premiering a new lineup of Echo smart-speaker and smart-display devices that work on the Alexa voice-driven home assistant platform.

These devices convey a lot of the aesthetics one would have seen in science-fiction material or “future living” material written from the 1950s to the 1970s. It is augmented by an indoor camera drone device that Amazon released around the same time.

As well, all of these devices have the spherical look that conveys that retro-futuristic industrial-design style that was put forward from the 1950s to the early 1970s like with the Hoover Constellation vacuum cleaner of the era or the Grundig Audiorama speakers that were initially designed in the 1970s thanks to the Space Race. You might as well even ask Alexa to pull up and play space-age bachelor-pad music from Spotify or Amazon Music through these speakers.

It is even augmented further with the base of the Echo and Echo Dot lighting up as a pin-stripe to indicate the device’s current status. This industrial design also permits the implementation of a 360-degree sound approach that can impress you a lot. It also is a smart-home hub that works with Zigbee, Bluetooth Low Energy and Amazon Sidewalk devices so you don’t need to use a separate hub for them.

Amazon Echo Dot Kids Edition that is available either as a panda or tiger – for the young or young at heart

The small Echo Dot comes in two different variants where one of these has a clock on the front while the other doesn’t. It also comes also as a “Kids’ Edition” with an option of a panda face or a cat face. It is offered as part of Amazon’s Alexa Kids program which provides a child-optimised package of features for this voice assistant. But I also wonder whether this can be ran as a regular Echo Dot device, which may appeal to those adults who are young at heart and want that mischievous look these devices have.

The Echo Show 10 smart display uses a microphone array and automatic panning to face the user. This is driven by machine vision driven by the camera and microphone array. But the camera has a shutter for your privacy. Of course you can use this device as a videophone thanks to the Alexa platform’s support for Amazon’s calling platforms, Zoom and Skype.

Amazon Echo Show 10 that can swivel towards you

What makes this generation of Echo devices more interesting is that they implement an edge-computing approach to improve sound quality and intelligibility when it comes to interacting with Alexa. This is even opening up ideas like natural-flow conversations with Alexa or allowing Alexa to participate as an extra person in a multiple-person conversation. It is furthering Amazon’s direction towards implementing ambient computing on their Alexa voice-driven assistant platform.

But Google was the first to implement this concept in a smart-speaker / voice-driven assistant use case. Here, they used it in the Nest Mini smart speaker to improve on the Google Assistant’s intelligibility of your commands.

Oh yeah, you can make and take Zoom or Skype videocalls on an Amazon Echo Show 10

I do see this as a major direction for smart-speaker and voice-driven-assistant technology due to improving responsiveness and user interaction with these devices. It may even be about keeping premises-local configurations and customisations on the device’s memory rather than on the cloud, which may improve a lot of use factors. For example, it may be about user privacy due to minimal user data held on remote servers. Or it could be about an optimised highly-responsive setup for the home-automation setups we build around these devices.

What needs to be looked at is a way to implement localised peer-to-peer sharing of data between smart speaker devices that are on the same platform and are installed within the same home network. This can allow users to have the same quality of experience no matter which device they use within the home.

There also has to be support for localised processing of data by devices with the edge-computing smarts for those devices that don’t have that kind of operation. This would be important if you bring in a newer device with this functionality and effectively “push down” the existing device to a secondary usage scenario. In this use case, having another device with the edge computing smarts on your home network and bound to your voice-driven-assistant platform account could assure the same kind of smooth user experience as using the new device directly.

These Amazon Echo devices are showing a new direction for voice-driven home assistant devices to allow for improved intelligibility and smoother operation.

Articles

The Google Nest Mini smart speaker – a follow on from the Home Mini smart speaker and having its own local processing to improve Google Assistant’s responsiveness

Google Nest Mini gets louder and gains onboard Assistant processing | SlashGear

Google debuts Nest Mini with wall mount and dedicated ML chip | VentureBeat

From the horse’s mouth

Google

Nest Mini

Nest Mini brings twice the bass and an upgraded Assistant (Product Blog Post)

My Comments

The Google Nest Mini smart speaker, which is the successor to the Google Home Mini smart speaker, shows up a significant number of improvements including a richer sound. But it has also come about with the idea of locally processing your voice commands for better Google Assistant performance.

The traditional approach to processing voice commands that are said to a smart speaker or similar device is for that device to send them out as a voice recording to the cloud servers that are part of the voice-driven assistant platform. These servers then implement their artificial-intelligence and machine-learning technology to strip background noise, interpret the commands and supply the appropriate replies or actions back to that device.

But Google has improved on this by using a leaf out of the book associated with edge-computing technology. This is where some of the data storage or processing is performed local to the user before the data is sent to a cloud computing system. Here, Google uses a dedicated machine-language processor chip in their Nest Mini smart speaker to do some of the command processing before sending data about the user’s command to the Google Assistant cloud system.

It reduces the idea of your Google Nest Mini smart speaker being a simple conduit between your home network and the Google Assistant cloud. The key benefit is that you see a quicker response from the Google Assistant via that device. You also have the benefit of reducing the Internet bandwidth associated with handling the voice-driven home assistant activity, avoiding reduced performance for online gaming or multimedia streaming.

Google is working on taking this further with having Google-Assistant-based devices that have this kind of local processing process logic associated with user requests and programmable actions locally. It also includes keeping the logic associated with the Assistant liaising with other smart devices local to your home network, allowing for improvements to performance, user privacy and data security.

It could be seen by Amazon and others as a way to improve the performance of their voice-driven home-assistant platforms. This is more so where the competition between these platforms becomes more keen. As well, there could be a chance for third-party Google Assistant (Home) implementations to look towards using local processing to improve the Assistant’s response.

An issue that will crop up is having multiple devices that have this kind of local processing existing on the same home network help each other to increase the voice-driven assistant’s performance. This can also include using a software approach to make the devices equipped with the local processing provide improved performance for those that don’t have this processing. It will be an issue with the likes of Google Nest Mini and similar entry-level devices that appeal to the idea of having many installed around the house, along with the idea of equipping smart displays with this kind of local processing.

What I see of this is that the use of edge-computing technology is coming to the fore as far as improving responsiveness in the common voice-driven home-assistant platforms.

QNAP TS-251 2-bay NAS – units like this could become a capable edge-computing device

The high-end network-attached storage system is a device able to augment the cloud computing trend in various forms. This is by becoming a local “edge processor” for the cloud-computing ecosystem and handling the data that is created or used by end-users local to it.

High-end network-attached-storage systems

We are seeing the rise of network-attached-storage subsystems that are capable of running as computers in their own right. These are typically high-end consumer or small-business devices offered by the likes of QNAP, Synology or NETGEAR ReadyNAS that have a large app store or software-developer community.

The desktop variants would be the size ranging form half a loaf of bread to a full bread loaf, with some rack-mounted units about the size of one or two pizza boxes.This is compared to servers that were the size of a traditional tower computer.

But some of the apps work alongside public cloud-driven online services as a client or “on-ramp” between these services and your local network. A typical use case is to synchronise files held on an online storage service with the local storage on the NAS unit.

These high-end network-attached-storage devices are effectively desktop computers in their own right, with some of them using silicon that wouldn’t look out of place with a traditional desktop computer. Some of these machines even support a “local console” with a display connection and USB connections that support keyboards and mice.

Cloud computing

Cloud computing takes an online-service approach to computing needs and, in a lot of cases, uses multiple computers in multiple data centres to perform the same computing task. This is typically to host the data in or close to the end-user’s country or to provide a level of scalability and fault-tolerance in the online service approach.

A cafe like this could benefit from big-business technology without paying a king’s ransom thanks to cloud computing

Small businesses are showing an interest in cloud-driven computing solutions as a way to come on board with the same things as the “big end of town” without paying a king’s ransom for hardware necessary for an on-premises computing solution. In some cases, it is also about using different endpoint types like mobile-platform tablets for daily use or as a management tool, underscoring such concepts as low cost or portability that some endpoints may offer.

Typically, this kind of computing is offered “as a service” where you subscribe to the service on a regular, usually monthly or annual, basis rather than you spending big on capital expenses to get it going.

But, due to its nature as an always-online service, cloud computing can cause reliability and service-availability issues if the Internet connection isn’t reliable or the service ends up being oversubscribed. This can range from real-time services suffering latency towards a cloud-computing experience becoming unresponsive or unavailable.

Then there is the issue of privacy, data security, service continuity and data sovereignty which can crop up if you change to a different service or the service you use collapses or changes hands. It can easily happen while cloud-computing faces points of reckoning and the industry goes in to consolidation.

Edge computing

But trends that are being investigated in relationship to the “Internet Of Things” and “Big Data” are the concepts  of “edge” and “fog” computing. It is based around the idea of computing devices local to the source or sink of the data that work with the locally-generated or locally-used data as part of submitting it to or fetching it from the cloud network.

It may allow a level of fault-tolerance for applications that demand high availability or permit scalability at the local level for the cloud-computing application. Some systems may even allow for packaging locally-used data in a locally-relevant form such as for online games to support local tournaments or an online movie service to provide a local storage of what is popular in the neighbourhood.

The ideas associated with “edge” and “fog” computing allow for the use of lightweight computer systems to do the localised or distributed processing, effectively aggregating these systems in to what is effectively a heavyweight computer system. It has been brought about with various early distributed-computing projects like SETI and Folding@Home to use personal computers to solve scientific problems.

What is serving the edge-computing needs

This Qarnot Q.Rad heater is actually a computer that is part of edge computing

Some applications like drones are using the on-device processing to do the local workload. Or we are seeing the likes of Qarnot developing edge-computing servers that heat your room or hot water with the waste heat these computing devices produce.  But Amazon and QNAP are working on an approach to use a small-office NAS as an edge-computing device especially for Internet-Of-Things applications.

The NAS serving this role

Here, it is about making use of these ubiquitous and commonly-available NAS units for this purpose as well as storing and serving data that a network needs. In some cases, it can be about the local processing and storing of this locally-generated / locally-used data then integrating the data with what is available on the cloud “backbone”.

For some applications, it could be about keeping enough data for local needs on the NAS to assure high availability. Or it could be about providing scalability by allowing the NAS to do some of the cloud workload associated with the locally-generated data before submitting it to the cloud.

The NETGEAR ReadyNAS on the right is an example of a NAS that is capable of being an edge-computing node

This may be of importance with IT systems that are moving from a totally on-premises approach towards the use of cloud-computing infrastructure with data being stored or processed online. It is where the focus of the cloud infrastructure is to make business-wide data available across a multi-site business or to provide “portable access” to business data. Here, a NAS could simply be equipped with the necessary software to be a smart “on-ramp” for this data.

For small and medium businesses who are moving towards multiple locations such as when a successful business buys another business in another area to increase their footprint, this technology may have some appeal. Here, it could be about doing some pre-processing for data local to the premises before submitting to the cloud as part of an online management-information-system for that small effort.  As well, it could be about keeping the business-wide data “in-sync” across the multiple locations, something that may be important with price lists or business-wide ledgers.

This kind of approach works well with the high-end NAS units if these units’ operating platforms allow third-party software developers to write software for these devices. It can then open up the possibilities for hybrid and “edge” computing applications that involve these devices and the network connectivity and on-device storage that they have.

Conclusion

What needs to happen is that the high-end network-attached-storage systems of the Synology or QNAP kind need to be considered as a hardware base for localised “edge computing” in an online “cloud-computing” setup.

This can be facilitated by the vendors inciting software development in this kind of context and encouraging people involved in online and cloud computing to support this goal especially for small-business computing.

Article 

Western Digital 14TB hard drive sets storage record | CNet

From the horse’s mouth

HGST by Western Digital

Ultrastar HS14 14Tb hard disk

Product Page

Press Release

My Comments

Western Digital had broken the record for data stored on a 3.5” hard disk by offering the HGST by WD UltraStar HS14 hard disk.

This 3.5” hard disk is capable of storing 14Tb of data and has been seen as a significant increase in data-density for disk-based mechanical data storage. It implements HelioSeal construction technology which yields a hermetically-sealed enclosure filled with helium that leads to thinner disks which also permit reduced cost, cooling requirements and power consumption.

At the moment, this hard disk is being pitched at heavy-duty enterprise, cloud and data-center computing applications rather than regular desktop or small-NAS applications. In this use case, I see that these ultra-high-capacity hard disks earn their keep would be localised data-processing applications where non-volatile secondary storage is an important part of the equation.

Such situations would include content-distribution networks such as the Netflix application or edge / fog computing applications where data has to be processed and held locally. Here, such applications that are dependent on relatively-small devices that can be installed close to where the data is created or consumed like telephone exchanges, street cabinets, or telecommunications rooms.

I would expect that this level of data-density will impact other hard disks and devices based on these hard disks. For example, applying it to the 2.5” hard-disk form factor could see these hard disks approaching 8Tb or more yielding highly capacious compact storage devices. Or that this same storage capacity is made available for hard drives that suit regular desktop computers and NAS units.

Article – French language / Langue Française

A small server in an apartment block’s basement communications room could be part of a distributed computing setup

Le phénomène Pokémon GO révéle nos besoins en datacenters de proximité | L’Usine Digitale

My Comments

This year has shown a few key situations that are placing traditional data-center technology under stress. This is based around fewer large data centers placed sparsely through a geographic area and used primarily to store or process data for many businesses.

One of these is the popularity of Pokemon GO. As people started playing this augmented reality game on their smartphones, there was the need to draw down data representing the app from the different data centers associated with the mobile platforms’ app stores. Then, during play, the game would be exchanging more data with the servers at the data centers that Niantic Labs controls. In some cases, there was problems with underperformance due to latency associated with this data transfer.

.. as could one of these Qarnot Q.rad room-heater servers

Then lately, there was a recent attack, purported to be a denial-of-service attack, against the data centers that were being used to collect the data for the census taking place in Australia on Tuesday 10 August. It is although the census is being targeted towards an online effort where households fill in Web pages with their data rather than filling out a larger book that is dropped off then collected.

Both these situations led to data-center computers associated with these tasks failing which effectively put a spanner in the works when it came to handling these online activities.

What is being shown is that there needs to be an emphasis on so-called “edge computing” or the use of small localised data centers also known as “cloudlets” to store and process data generated in or called upon by a particular area like a suburb or an apartment block. These data centers would be linked to each other to spread the load and pass data to similar centers that need the data.

One application that Netflix put forward was their “Open Connect Appliance” which as a storage device that an ISP or telco could install in their equipment rack if they end up with significant Netflix traffic. This box caches the local Netflix library and is updated as newer content comes on line and older locally-untouched content is removed. Such a concept could be taken further with various content delivery networks like Cloudflare or those implemented by the popular news services or social networks.

The trend I would initially see would be suburban telephone exchanges, cable-TV headends or similar facilities being seen as a starting point for local micro datacenters that serve a town or suburb. Then this could be evolving to street cabinets associated with traffic signals, FTTC/FTTN services and the like, or the basement telecommunications rooms in multi-tenancy buildings being used for this purpose with these smaller data centers being used to serve their immediate local areas.

Qarnot, with its Q.Rad room heaters that are actually data servers, weighed in on the idea that a cluster of these room heaters in a premises or town could become effectively a local “micro data center”.

As for applications that I would see for these “micro data centers” that implement region-focused data processing, these could include: distributed content delivery of the Cloudflare or Netflix kind; localised “store and process” for data loads like a nation’s census;,  online gaming of the Pokemon GO kind; and distributed-computer applications that ask for high fault-tolerance. There will still be the need for the “virtual supercomputer” that would be needed for huge calculation loads like sophisticated financial calculations or 3D animation renderings which a collection of these “micro data centers” could become.

Similarly, the issue of distributed localised computing concepts like edge computing and local “micro data centers” could reduce the need for creating large data centers just for handling consumer-facing data.

What could be seen as affecting the direction for cloud-based computing would be the implementation of localised processing and storage in smaller spaces rather than the creation of large data centers.