When I was a kid the year 2020 sounded impossibly distant. In fact, if you grew up in the 1970's any year from 2000 on seemed futuristic; the stuff of pure science fiction. I figured surely by 2020 we'll have colonized Mars and extended our life spans to 150 years, right?
Well, not so much, but 2020 will bring us one thing we don't have now: 5G networking (if all goes well).
What is 5G networking?
5G is the next level of mobile networking; the successor to 4G. As a background, current 4G networks run at about 100 Megabits per second while 5G will travel at 10 Gigabits per second - theoretically 100 times faster, but in real life we'll probably see about two-thirds of that boost. It's still nothing to sneeze at, especially when you consider that 3G networks run at a paltry 384 Kilobits per second.
As with any increase in network speed and capacity, companies have big plans for how to use these new capabilities. Obviously, upload and download speeds will drop to a fraction of their current level. Streaming audio/video will work much more efficiently. All sorts of new infrastructure related to the Internet of Things (IoT) and advances in environmental or vehicular controls are being eagerly mapped out to make the best use of the increased speeds associated with 5G.
But, to paraphrase that line from "Spider-Man" about great power and great responsibility (sorry, even I think it's been used to death) along with great opportunity comes a greater potential for error. And that's where it's crucial to understand the challenges in advance.
How can we prepare for 5G?
In order to prepare for the new 5G standard it's important for mobile operators to look at how their support and architectures will accommodate it and to find potholes that might impact service availability that much faster. The situation is made more challenging by the fact that current 3G/4G methods for monitoring and managing networking incidents won't scale for 5G environments, especially those which are rapidly changing.
Deutsche Telekom is one of the leading mobile operators driving 5G advancement. They have launched a 5G innovation lab (called 5G:haus) to promote further advancement and standardization of 5G mobile technology as well as to provide a platform to engage with customers throughout the 5G development process. Their goal is to get companies of all sizes and backgrounds involved to explore, debate, test and brainstorm new technologies which will fulfill the potential of 5G.
According to ZDNet, "The 5G:haus will allow us, in cooperation with leading partners, to develop the architecture, to foster innovation and to steer standardization work. The lab will also provide a platform to engage with our customers at a very early stage of 5G development," says Bruno Jacobfeuerborn, chief technology officer for DT." Subsequently, Deutsche Telekom has selected specific companies to participate. One such company is called Moogsoft.
What is Moogsoft doing with 5G?
Moogsoft is a service assurance vendor - the only one announced (as of present day) by Deutsche Telekom to participate within their 5G:haus lab. Their perspective is that 5G networks are too complex to assure services using traditional approaches. Therefore, a big data, real-time unsupervised machine learning approach to anomaly detection which can push notify occurrences of incidents and their impact on 5G services is the only way to assure a smooth transition to 5G. Their product, Incident.MOOG, is a part of this strategy.
According to Marketwired.com, "Incident.MOOG is designed to address the rapidly-changing IT environments that telcos, web-tech enterprises, financial institutions, and managed service providers face as they scale to address new markets, optimize their infrastructures and differentiate with superior service quality... Moogsoft enables deep monitoring of both new and old IT infrastructure components like switches, access gateways, controllers, servers, cloud virtualization software, as well as domain-specific monitoring tools, to deliver a single source of truth for complex, service-affecting issues in the network."
Furthermore, "Incident.MOOG detects anomalies in real time to reveal service-affecting incidents as they unfold, dramatically reducing Mean-Time-to-Repair (MTTR), and enables knowledge capture and reuse through collaborative 'Situation Rooms.'"
Moogsoft describes Incident.MOOG, as "an IT incident early warning and collaborative remediation system used by web-scale enterprises and service providers within IT Operations Management (ITOM), IT Service Management (ITSM) and DevOps teams. Using patent-pending machine learning technology, Incident.MOOG automates the early detection of service failures, reduces the noise without dependence on preset rules or models, creates contextualized situations of related alerts, and then facilitates collaboration via its Situation Rooms, either averting business disruption or orchestrating the rapid restoration of business services."
So what does this have to do with 5G?
"5G is just one of the multiple technology transformations in this new era that is forcing mobile operators to revamp how they detect and remediate complex system problems," said Phil Tee, CEO of Moogsoft. "The telco industry is known for the complex IT infrastructures resulting from multiple acquisitions and the rapid advancement of technology solutions. Reducing time-to-remediation for service-affecting issues is critical in this competitive market. The only way to support these new, dynamic infrastructures is with a next-generation, agile manager-of-managers toolset that automates real-time anomaly detection across all layers and domains."
I had the opportunity to chat with network engineers at Moogsoft about their approach to 5G.
Scott Matteson: "What can you tell me about 5G networks that will entice my readers?"
Moogsoft: "Gartner forecasts that the number of networked devices will reach 25 billion by 2020, and IDC predicts that it will become a $3 trillion market by then. Considering the predicted boom in the Internet of Things, with sensors embedded into nearly every gadget and appliance and the intercommunication that it will demand, 5G will have a central role in this next generation of networking. And telecom providers will be seeking new communication infrastructure to support this data load: radio frequencies for greater throughput, lower latency and higher connection density. These frequencies are currently underutilized for smart technology ranging from in-home appliances to city-wide street lights.
The infrastructure to deliver 5G Next Generation Mobile Networks will be very similar to that utilized for Private Cloud and NfV based Virtualized Managed Services. Expect high utilization of virtualized computing fabric - probably OpenStack (if I were a betting man!). This really is the convergence of Mobile Networking and Cloud Computing services.
In addition, extensive use of NfV is to be expected too - 5G in reality will offer the first ubiquitous high speed roaming - and so will offer businesses the ability to truly utilize the promise of location independent low latency communications. Telepresence and group working from personal devices will be viable with this speed and bandwidth - so expect 5G to enable the mobile operators to truly compete with the wired/broadband networking companies for Business oriented services.
The provision of this level of bandwidth and network speed will also allow the 5G service providers to offer their own Public and Private Cloud Services - after all - why go off their network to an AWS/Google or Azure which will increase latency to deliver an optimum application experience.
Since 5G is all IPv6 based, 5G enables IoT scale numbers of devices and IoT scale data volume harvesting, processing and Groupthink content delivery based upon that processed data back to the IoT devices at the same time.
The business opportunities that evolve from this quantity and [almost] limitless data types (more than we can think of today) offers some new propositions to evolve - both on the positive side and the negative side. One example: 5G heralds the real end of privacy."
SM: "What is the Deutsche Telekom 5G innovation lab intended for; which problems does it seek to solve, how will solutions be applied?"
Moogsoft: "The 5G Innovation Lab's focus is on creating industry standards and interoperability between networks. It closely works with the Next Generation Mobile Networks (NGMN) Alliance to streamline end-to-end operator requirements across time and geographies. For the end user, the focus is on engaging customers at the onset and ensuring a fluid customer experience for the ever-growing use of multiple devices. For enterprises, the focus will be on cost, flexibility, accessibility, scalability and room for innovation across a unified ecosystem.
Deployment and production customer usage of 5G networks is 5 years away. DT has announced that they intend to be a leader, not a follower, in 5G Next Generation Mobile Networking and the service offerings layered onto 5G, so we would expect that they will use the lab to explore scenarios and the issues that arise. One of the useful things about Moogsoft's real-time anomaly detection approach is that it addresses "The Donald Rumsfeld Problem." Moogsoft has a patent relating to the ability to detect 'unknown unknowns'; that is, Incident.MOOG is able to infer unusual behavior from streaming alert data without one having to pre-describe or learn that a given behavior is unusual. Modern IT and telecommunications fabric is like quicksand - being virtualized, the network and computing fabric can change at any time. This means it is impossible to model all fault and impact conditions - there could be an infinite set of scenarios. Incident.MOOG is able to infer fault and impact conditions without models. So in a 5G networking world, where we do not know what we do not know, using Incident.MOOG to help detect unusual behavior is of high value."
SM: "Can you break the description of Incident.MOOG down into laymen's terms? I have a diverse reader base, from technologists to executives, and want to ensure that I get the capabilities and details down properly."
Moogsoft: "Faults happen. Things break. When you can see the break in a cable, you know that the users of either end of the cable will be impacted.
If you see two cables both connecting to the same two parties, if you notice one cable has a break, you know the two parties are still connected.
When you have an unknown number of cables connecting an unknown number of parties together but each cable may be able to be switched between parties, it is very hard to work out which cables have a break when some parties are connectivity is interrupted. Incident.MOOG will help you work out which cables have breaks and at any given time, which parties are unable to connect.
In other words, without knowledge of how something works, Incident.MOOG is able to detect that there is a problem in real-time, indicate what is impacted by that problem, and, inform the people who need to work on it in order for them to diagnose and resolve the issue more quickly. Incident.MOOG also allows multiple teams who may be involved in the diagnosis and resolution of an issue to be situation aware and work collaboratively together. A problem shared is a problem halved after all!
Incident.MOOG is an IT event monitoring platform that helps detect, diagnose and then remediate an issue before it can negatively impact a critical technical or business process. For example, some businesses may not know that an application or service may be "down" until they get a complaint from a customer. Incident.MOOG pulls data from a variety of network sources, IT tools and even Twitter to identify early warning signs of issues and then notify the relevant IT and development staff to address it. The goal is to identify issues before they affect service availability and impact the business."
SM: "Can you elaborate on how Moogsoft will stream big data as part of this lab; what effects are intended and how it will help monitor 5G networks?"
Moogsoft: "Incident.MOOG does not stream data. Incident.MOOG receives status messages from equipment (network, computers, hypervisors, applications, NfV's, customer sentiment Twitter streams...any textual data which indicates state or opinion) - in the case of computers, Syslog and SNMP, in the case of Applications, Log4J, etc. - and as those messages are created our patented layers of 'Cleaning', 'Contextualizing' and Collaborating' algorithms go to work. So, Incident.MOOG can take status telemetry from any source and infer relationships with other sources.
The move to virtualized infrastructures and users' adoption of mobile applications in the IT world has increased raw event and log data rates by a significant factor, overwhelming operations and support staff. This is just a preview to the burden that 5G network operations support will experience.
Traditional monitoring tools and services that were designed for second and third-generation mobile networks cannot scale to address the complexity and load of a 5G-era environment. Through the use of streaming big data from throughout the network, Moogsoft is enabling Deutsche Telekom to uncap their event/log telemetry data and rate of infrastructure change, while at the same time detecting faults and the impact of those faults earlier, all without increases in operations and support resources. As the amount of data that needs to be monitored rapidly grows, the IT team at Deutsche Telekom will be empowered with the monitoring tools they need to scale operations effectively. "
SM: "What does Moogsoft bring to the table that other vendors don't?"
Moogsoft: "In the first place, Incident.MOOG uses a patented unsupervised machine learning technique to detect unusual behavior (or anomalies) without the need to pre-describe or have seen the anomaly before in order to monitor and detect it.
Fundamentally, if you think this through, if a virtualized elastic computing based infrastructure and applications can change at any time, there could be an infinite set of failure / impact scenarios. With other tools, you either need to have seen the issue / behavior previously in order to monitor it in the future, or you need to wait until your customers call you to tell you they have an issue with your service. Neither is very good for the customer.
Moog will detect things before they become service impacting and 'push notify you' with a warning. This means you can avert disruption - because Moog will detect things that you do not know about.
Secondly, Incident.MOOG represents / presents the anomaly as a cluster of related alerts. This means Moog offers a narrative of the failure / impact. So, operations are notified early then able to quickly diagnose the causality of the issue with reduced research effort.
Third, because Incident.MOOG captures the resolution activities as knowledge, when similar situations are detected in the future, that knowledge is shared with the operators, enabling them to more quickly resolve the issue.
These three capabilities are unique to Moogsoft.
How does this relate to customer value?
- 1. In an enterprise customer example, Incident.MOOG has demonstrated a reduction in Trouble Ticket work items for Applications / DevOps support teams by greater than 86%.
- 2. For another customer (a Web/Mobile Tech Service Provider), Incident.MOOG has reduced the number of actionable work items for Tier 1 support people by greater than 75%.
- 3. Incident.MOOG has reduced the number of escalations from Tier 1 support to Tier 2 by greater than 90%.
- 4. For another customer, Incident.MOOG has enabled the same number of operators staff to handle a greater than 5x increase in workload.
This business value is unique to Incident.MOOG.
Moogsoft has designed Incident.MOOG for high scalability. Although the company is only 3 years old, the customer base combined is using Incident.MOOG to assure the availability of greater than 2 million network, compute and storage devices, virtual computers and applications.
Moogsoft's machine learning and socialized workflow approach provides the agility, visibility and intelligence to gather pertinent information from across an application stack (across multiple hybrid environments and third-party tools) to triage possible service affecting events before they impact network performance. Legacy monitoring tools rely on rules that simply don't apply to dynamic, ever-changing IT infrastructure. Incident.MOOG's machine learning algorithms "see" changes in the network as they occur and adjust to reflect that, ensuring that the alerts and alarms the system identifies are relevant."
I appreciate the insights provided by Moogsoft's team. It will be interesting to see the results of the 5G:haus lab; the opportunities presented and the foundations laid to prepare for the advent of 5G. More to come on this topic, I'm sure!
Scott Matteson is a senior systems administrator and freelance technical writer who also performs consulting work for small organizations. He resides in the Greater Boston area with his wife and three children.