Connect with us


Ciena uses machine learning to heal the scars, horror of network management



Does machine learning still need people?
Vijay Raghavan, executive vice president and chief technology officer Risk and Business Analytics for The RELX Group, talks with Tonya Hall about a balance between analytics and intuition.

“The scars” and “that horrible world” are some of the terms for network management, according to one who’s been in the trenches. 

Kailem Anderson was with Cisco Systems for 12 years prior to joining fiber-optics giant Ciena last year. As vice president of portfolio and engineering for the Blue Planet, a software division of Ciena, he is trying to help avoid such pain for those who must keep networks running. 

“I managed customer networks, and I spent a lot of time hiring analysts to watch the network, to watch alarms, and to build big strings of rules,” for networking monitoring, says Anderson. His breezy Aussie accent gives a certain lightness to what sounds like a rather miserable affair.

At $26 million in revenue in 2018, Blue Planet was a tiny fraction of Ciena’s roughly $200 million in software revenue in 2018 and $3 billion in total revenue. But it increased by a healthy 66%, and it can bring higher profit margin than Ciena’s optical networking equipment sale. It also offers the company a recurring revenue stream that is highly appreciated by Wall Street. Those economic aspects, plus the fact that it can be strategic in designing customers’ networks, make it an important part of where Ciena is headed as a company.   

Also: Is Google’s Snorkel DryBell the future of enterprise data management?

Figuring out what’s gone wrong in a network involves detective work at several levels of what’s known as the “stack” of protocols, the Open Systems Interconnect, or “OSI.” Some information comes from the bottom of the stack, if you will, the “layer one,” which consists of the physical medium of transmission. That could be, for example, coaxial cabling or fiber-optic links. 

At the next layer above that, layer two, raw bits are packaged into bundles, such as Ethernet frames, and there’s all kinds of information to be gleaned about the state of those frames of data as they move through the fibers and cables of the network. The next layer up is layer three, where data is packaged as Internet-addressable packets, again, with lots of their owing information to be gleaned, such as routing and switching information about where the packets are going. 

From there, one can go on up to higher levels, layers four through seven, the domain of applications, and get information about who an individual application is placing its data into those internet packets and whether it is having any trouble doing so. 

Take the example where there is an transponder failure on one of two optical links. That leads to a route change in the multi-protocol label system, or MPLS. The network equipment reports congestion along the IP route as a link shoulders the burden of more traffic, and an end user experiences heavy delays using the network. All these are part of the same problem, Anderson explains, but getting from the user experience to the transponder failure can be a mystery. 

Traditionally, a systems administrator sees the various items in a disparate fashion, with signals at each of the OSI layers coming from different telemetry systems, such as SNMP monitors, the systems log, a third thing that tracks “flows,” and then information coming from an individual piece of equipment, such as information about a recent configuration change — none of which are coordinated. 

What looks like bad user performance from one angle looks like an MPLS routing issue or an IP bandwidth issue at another level, leading to a serious piece of detective work to find the culprit, the transponder failure. 

Also: Google Brain, Microsoft plumb the mysteries of networks with AI

A ticket gets created, and it ping-pongs between teams, with no one team having visibility into the other side, says Anderson. “Eventually they solve it, they have engineers inspect the matter, but it’s very inefficient.”

Sys admins must try and construct systems of rules as to what every possible combination of factors could mean. “They spend 1,000s of hours building these rules,” says Anderson. “It’s a zero sum game to spend that time to identify all the different scenarios.”

Instead, Blue Planet tools can train the network software using a combination of labeled examples, known as supervised learning and reinforcement learning, where the computer explores states of affairs and possible next steps. 

With that combination, the software can be trained to identify patterns “up and down the stack” that are difficult to piece together with a rules-based system. 

“We want to have the system learn to identify those scenarios, to basically help us get to the root cause much more quickly, and to use that information to close the loop,” he says, and then have a supervisor come into the picture only once that outline has been determined. 

Also: Intel-backed startup Nyansa chases the total problem in the AI of network monitoring

The tools necessary to do this are mostly starting from off-the-shelf machine learning models, says Anderson. “Most of this, yes, we can get from the cloud guys,” he says, referring to the various enterprise-grade machine learning offerings in cloud computing facilities. “We use them all,” though the tools can also be run solely on-prem. “It’s six and one half dozen of the other at the moment, but I think analytics is ultimately a good thing to move into the cloud.”

Open-source tools such as SparkML play a big role in organizing all the telemetry data. 

The technology of machine learning, says Anderson, has matured substantially in recent years to make the investment in labeling network events pay off. 

“Five years ago I was playing with this and with the amount of effort that needed to go into labeling, the risk versus value I was getting was questionable,” he says. “With the hardening of the algorithm, and the maturity of AI, that effort-to-reward ratio has compressed significantly. You only have to do a reasonable amount of tagging now and the outputs are significant.”

Anderson maintains there is another dimension in the shift to machine learning, which is that a more comprehensive sense of the network emerges that may lead to different ways or structuring and maintaining networks. 

Traditionally, many sys admins will simply turn off sources of information, says Anderson, which is understandable, because of the information overload, but it means that network administrators are throwing away important clues. 

“That’s the complexity in operating with a million different data sources,” he observes. “The traditional way to mange an operations team is to filter the information, almost turn off the information that is too much.

“At Cisco, if I was running a service provider network, I would get in the vicinity of a million events a day, and I might have an operations team of 40 to 50 people who have to handle all that.”

As a consequence, admins end up only looking for “what they deem fair scenarios,” and “are turning off performance-based scenarios,” information about the relative quality of the network. 

But, says Anderson, “you don’t want to turn off the information, you want to funnel it, and use it to identify what conditions are driving consistent scenarios,

Must read

“Eventually, solutions could be different if they’re trained,” he offers. Data may lead to structuring things differently. “Usually, you have a planned network condition, but then an actual network condition; through learning, you might find the actual is more optimal than planned, and then execute a policy” based on that new insight. 

There are new frontiers to achieve, such as delivering analysis of the data in a “graph database” format, says Anderson. “We are in the operations and network world, and so you want to visualize all this in a network graph concept.” Some customers “want to see it just programmatically propagate to northbound systems that are going to leverage that information, to be able to visualize with a graph database and have APIs to send that northbound information to the BSS layer.”

The one catch at the moment in all this is that systems administrators are not yet ready to close the loop, so to speak, and let machine learning completely take over and automate both detection and resolution of network issues. 

“This isn’t a tech limit, it’s a cultural aspect,” he says. Machine learning systems are probabilistic, not deterministic. Hence, while they can detect many failure issues, there is a reluctance to automate what could be a false positive scenario. “You only need to screw up .0001% of the time and that’s a big issue.”

“I still think we are a little bit away in terms of closing the loop, I think it’s trust in the technology. It will happen incrementally, where you can close the loop on something non-catastrophic, that doesn’t create a failure scenario, where there is low risk, and then other areas over time

Source link

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *


Biden EV plan “the largest mobilization of public investment” since WW2



The US government plans to replace its fleet of vehicles with electric alternatives, part of a huge public investment in equipment that President Biden says will be the largest since World War II. The goal was announced today, as Biden discussed the Buy American Executive Order and strategies to strengthen American manufacturing and create jobs in the US.

“The federal government also owns an enormous fleet of vehicles,” Biden said during the press conference, “which we’re going to replace with clean electric vehicles made right here in America, creating millions of jobs, a million autoworker jobs, and clean energy, and vehicles that are net-zero emissions. And together this will be the largest mobilization of public investment in procurement, infrastructure, and R&D since World War II.”

Certainly, there’s no shortage of vehicles being used across the government’s various departments. Each year, those departments are required to submit records on owned, leased, and commercially leased vehicles in their fleets. For 2019, civilian agencies owned more than 158,000 vehicles, while military agencies owned more than 62,000 vehicles.

The biggest fleet, however, is used by the US Postal Service. In the 2019 figures, it reported owning more than 224,000 vehicles. Tallied across all the agencies, there’s more than 445,000 owned vehicles on the books, and more than 200,000 leased in some form, with total costs of around $4.4 billion.

Typically, domestic vehicles already far outweigh the use of foreign passenger vehicles and trucks in use by the US government. Indeed, of the roughly 645,000 strong fleet in 2019, only around 6-percent were foreign-made. However, should even a small percentage be replaced with electrified vehicles, that could represent a huge shift in emissions.

The current reporting does not break down the vehicles by drivetrain type – beyond a separate category for low-speed electric vehicles (LSEV) – so it’s unclear how many might already be battery-electric or hybrids.

Still, it’s only in recent years that there have been viable options for replacing mainstream vehicles with zero-emission alternatives, and even then some categories are still awaiting production EVs. Ford and Chevrolet are both preparing electric pickup trucks, as is Tesla, and several American automakers are working on electric SUVs and sedans.

Startups like Rivian and Canoo are developing both passenger cars and SUVs and electric delivery vehicles, while Ford has an e-Transit in the pipeline, an all-electric version of its best-selling van. Earlier this month, GM announced a new brand, BrightDrop, to focus on electric logistics.

“The dollars the federal government spends on goods and services are a powerful tool to support American workers and manufacturers,” the White House said today. “Contracting alone accounts for nearly $600 billion in federal spending. Federal law requires government agencies to give preferences to American firms, however, these preferences have not always been implemented consistently or effectively.”

Full details of today’s executive order are yet to be published in the Federal Register by the government.

Continue Reading


Arcimoto is planning a tilting electric three-wheeler and it sounds epic



Get ready for the electric vehicle market to lean over, with Arcimoto announcing it’s acquired a tilting trike specialist to use the tech in future EV three-wheelers. Based in Oregon, Arcimoto currently offers all-electric vehicles for public and business use, with its FUV – or “Fun Utility Vehicle” – available to preorder from around $18k.

Today’s deal sees Arcimoto snap up Tilting Motor Works, which offers a leaning kit for motorcycles. Its TRiO system can adapt existing bikes, promising to keep their natural lean but adding stability with a second front wheel. It’s currently offered for Harley-Davidson, Honda, and Indian models, priced at $14k plus installation.

Now, TRiO will be used for future Arcimoto trikes. The system will allow the EVs to lean into corners for more engaging driving dynamics, as well as lock upright when at a stop. The extra front wheel aids in traction too, Tilting Motor Works says, as well as improving braking; the company will continue to offer its kits for traditional motorcycles as well.

Arcimoto’s current FUV supports two occupants, sitting one behind the other. It has a 75 mph top speed from dual electric motors, and a range of 102 miles of urban driving; the doors are detachable, and rather than a steering wheel inside there are handlebars with heated grips. Currently, the company is taking preorders in Florida and on the west coast of the US.

It’s not the only vehicle Arcimoto has in mind, however. Late in 2020, the company showed off its latest prototype, a three-wheeler it called the ROADSTER. Roofless, and with a chopped-down windshield, it promises a more traditional take on the electric trike segment.

Electric drive and three wheels is arguably where some of the most interesting experiments are taking place in mobility right now. Last month, for example, we took the ElectraMeccanica Solo EV out for a spin, a single-seat electric trike that aims to reboot commuting. Based on the fact that almost 90-percent of Americans typically drive alone, it trades cabin space and seating for much cheaper running costs.

What it doesn’t do, however, is tilt. For an example of that, you have to look to something like Toyota’s i-ROAD, a distinctive electric three-wheeler that could lean into corners according to how aggressively you steered. Offered in select locations in Japan, and part of Toyota’s aggressive electrification push for the next few years, the i-ROAD was never officially offered in the US.

Continue Reading


Lotus teases sports car future as Elise, Exige and Evora face the axe



Lotus is preparing a huge shake-up, with the iconic British sports car company confirming it’s axing three of its most memorable models to pave way for an all-new line-up. 2021 will be the end of the line for the Evora, Exige, and Elise, Lotus said today, dropping a new teaser about just what is intended to take their place.

We’ve already seen one element of that plan, in the shape of the beastly Evija hypercar. Announced in mid-2019, it promised to tap electrification for its potency, something Lotus demonstrated in action at its public debut in October last year.

Not everyone will be able to afford – or even find a spot on the waiting list for – a $2m+ EV behemoth. For that audience, Lotus has confirmed a new series sports car range, with the Lotus Type 131 expected to go into prototype production later this year.

It’ll be built at the automaker’s Hethel, Norfolk facility in the UK, which will undergo a $127m+ investment and see around 250 more engineers and manufacturing recruits added to the payroll. Helping foot the bill are shareholders Geely and Etika, which took ownership of Lotus in September 2017.

You can’t accuse Lotus of not making the most of its outgoing range. The first-generation Elise made its debut all the way back in 1995, a new model for Lotus but epitomizing its ethos of reducing weight in the name of increasing engagement. Though never the most powerful sport car, it nonetheless carved out a lingering reputation across several generations for its purity behind the wheel.

The Exige, meanwhile, arrived in 2000. A coupe to the Elise’s convertible, it built on the platform with race-focused technology to maximize performance. Come 2008, the Evora gave Lotus an entrant for the super-sports sector, tempering some of the automaker’s notorious focus on paring back creature comforts with a more luxurious, GT-minded approach.

The Type 131 range – part of what Lotus is calling its Vision80 strategy – will include at least three new models, replacing the Elise, Exige, and Evora. “Our renowned team of engineers, designers and technicians who are working on the new cars are acutely aware of the legacy from the Elise, Exige and Evora,” Matt Windle, executive director of engineering at the automaker, says. “Indeed, many were around when Elise was being developed.”

Earlier in 2020, rumors suggested Lotus could reboot the classic Esprit name for a new model. The new Esprit, it was hinted, could use a hybrid V6 powertrain, combining gas and electric power, though unlike James Bond’s car it would be unlikely to turn into a submarine.

Continue Reading