Emerging Tech to Help Guard Against the Malevolence of Cloud Outages
From the outermost edge of extreme networks to the center of the energy vortex, emerging technologies are set to permanently end cloud outages. Will they eventually eradicate the clouds, too?
![Rainy night with thunderstorm and lightning over The Gargoyles of Notre Dame in Paris Rainy night with thunderstorm and lightning over The Gargoyles of Notre Dame in Paris](https://eu-images.contentstack.com/v3/assets/blt69509c9116440be8/blte38d7d9e806cad21/64bf4e937b923b602267ed37/00Gargoyle-_Alexey_Panferov-alamy.jpg?width=700&auto=webp&quality=80&disable=upscale)
Rainy night with thunderstorm and lightning over The Gargoyles of Notre Dame in ParisAlexey Panferov via Alamy Stock
According to a recent Uptime Institute report, cloud outages are becoming more frequent. This problem is not a matter of mere inconvenience but of things far more concerning.
“Traditionally cloud outages were related to simple equipment or technical glitches,” says Talal Shamoon, CEO of Intertrust, a data trust and device identity platform provider. “But in a world of distributed data sets spanning multiple clouds with differing technologies geared at not interoperating, apps running in and across clouds written by different app providers and IoT hardware feeding sensor data to clouds, we have a Pandora’s box of issues that can make cloud outages not only more frequent but far more dangerous to companies.”
There’s no going back to simpler cloud days. The only choice is to innovate or move on to another job.
“In truth, few CIOs live to see this day [recovery after a cloud outage] and it’s partially for this reason that CIOs have one of the highest turnover rates within the C-suite,” says Ori Weizman, solution architect at Silk, a data migration platform for the cloud. “My advice is simple, if you’re given a second chance in an increasingly one-strike ballgame, don’t cut corners. Invest what you need to because an outage will always cost more down the line.”
Here's a heads up on emerging tech that aims to help you circumvent or instantly recover from cloud outages. You may want to investigate them further before finalizing your mid-term cloud management strategy.
Click through but be prepared for a few interesting surprises.
The term mobile edge computing morphed into multi-access computing somewhere along the way, but not everyone agrees the two terms mean the same thing. “While multi-access edge computing refers to the edge of any network, mobile edge computing refers to the edge of only the mobile network,” according to Verizon. Others agree with Verizon’s distinction in definitions.
“Mobile Edge Computing (MEC) is an emerging approach that locates data centers close to cellular towers and radio networks, improving resiliency and performance, especially for ultra-low latency mobile applications such as those made possible by 5G,” says Pat Wilbur, CTO and co-founder of Hologram.io, a SIM-enabled constant connectivity platform for IoT devices.
By either name, like all technologies, this too comes with a downside.
“The problem with MEC, or any cloud disaggregation for that matter, is that it extends your circle of trust from a few cloud providers to multiple network providers or even many data centers within a single network. Traditionally, data is encrypted in transit but then must be decrypted in order for the data center to perform computation. It’s one thing to entrust a few select data centers with your decrypted data, but not dozens, hundreds, or more,” Wilbur adds.
The model of edge computing itself is evolving. In this evolutionary round, the cloud can be circumvented entirely rather than augmented or partially replaced by common MEC schemes.
“A distributed model must be in place to create technology that’s not at risk during a cloud outage,” says Rob Spalding, CEO of SEMPRE, a micro-edge datacenter manufacturer. “First, the significant processing must occur in a node at the true edge, physically close to the user, reducing dependence on the need for a long connection back to a traditional datacenter. This is true edge processing rather than data collection, which is transferred back to the cloud. Second, multiple redundant paths must be in place to connect the user to the node, meaning alternatives to fiber such as free space optical or MM Wave [millimeter wave] are used.”
Both MM Wave of the 5G trend and free space optical networks are themselves new technologies, if more in application than by discovery or invention date. Most people have heard a lot about 5G through a barrage of ads and carrier offerings.
Fewer are familiar with free-space optics (FSO), aka free-space photonics (FSP), meaning broadband communications networks built of modulated visible or infrared (IR) beams of light. These are most often communications over laser light, but LEDs and IR-emitting diodes can also do the trick.
“The net reduction of an outage is easily understood with an example of a manufacturing plant that is highly automated and leveraging IoT to optimize production,” says Jeff Sieracki, senior director of product development at Lumen Technologies. “If compute is adjacent to the factory, then cloud or network outages will not stop production. Additionally, the proximity allows for fast analysis of data -- in near real-time -- for greater productivity.”
Gaps between multi-cloud apps restrict how well they perform and whether they perform in other clouds. Yes, we’re talking smart app integration to fill the cracks that workloads might fall through during an outage thereby greatly delaying their restart time.
“I expect we will see more solutions emerge that integrate a multitude of application and microservice testing solutions that help organizations identify architectural gaps in applications that operate in a multi-cloud environment -- and thus enable those applications to operate without interruption during an outage with a particular cloud provider,” says Tim Potter, principal at Deloitte Consulting.
This idea is heating up. “Increasingly, enterprises are looking to run applications across multiple clouds to increase their resilience to a cloud provider failure or outage,” he says.
Cloud-hopping apps are the next innovative idea in our lineup. Cool idea, huh?
“The seemingly straightforward way to build cloud outage resiliency is to have redundancy for your applications across cloud providers and regions, but this has always been too difficult and expensive for all but the most critical systems,” says Jye Cusch, co-founder of Nitric, a cloud-native app framework provider. “We’re seeing really interesting emerging technologies that make this newly possible for companies.”
Emerging code tools, such as Dapr and Nitric, are examples of new software development tools that support code portability and cloud-agnostic architecture.
Cusch says resilience is amplified when combined with a real-time data-sharing platform, like Vendia, to enable an on-demand pivot after an outage to another cloud, and/or the option to affordably run in multiple regions or cloud providers from the beginning. Distributed databases like Vitess and Temporal, are useful too, he says, as they help teams build distributed, highly resilient workflows.
“With these new tools, companies can now easily run in multiple regions and providers at all times, but most are still wary of the cost of keeping data synchronized due to data egress charges. For a lot of companies, the powerful resiliency tactic with these emerging tools is to keep their applications active in a single region or provider, with the ability to migrate on-demand when an outage occurs. This helps reduce total cost for companies that can weather the shorter outages and are mostly concerned about being able to recover from multi-day ones," Cusch adds.
Automation and AI aren’t new technologies per se, but some of their applied uses are just now emerging. Monitoring and managing workloads across clouds to ensure any cloud’s outage results in zero interruption is a big one. AI can anticipate any outage issue and take automated action to avoid outages or leaks.
“It is the only way a large organization can respond sufficiently to disruptions across what is likely to be thousands of diverse cloud workloads,” says Jeff Northrop, senior director, security center of excellence at Mars Inc., an American multinational manufacturer of confectionery, pet food, and other food products as well as a provider of animal care services. “Current best practices encourage efficient deployment models, high availability architectures, infrastructure as code, off-line backups, and out-of-band health monitoring. With those in place an organization can leverage those practices to automatically respond to events.”
“The industry is starting to offer solutions in this area, but more capabilities need to be developed, particularly when you consider what is necessary to respond to a critical cloud native service or SaaS going down," Northrop adds.
Here’s a different spin on old tech to add faster bounce to your bounce-back plans for dealing with a cloud outage.
“While object storage itself has been around for 20+ years, only now are we seeing the consideration and optimization of object storage for backup and recovery,” explains Tony Liau, vice president of product marketing at Object First, a provider of object-based storage. “The optimized performance enables the ability to run failed workloads directly from backup, ensuring companies are business as usual within minutes during any outages.”
Specifically, this innovation is a peer-to-peer database that enables devices to share data directly, meaning with no internet connectivity involved at all.
“The first wave of mobile devices was very weak, so they had to connect to the cloud or server and have the server do all the work,” says Adam Fish, CEO of Ditto, a company which has created an architecture that allows data to pass directly through devices in real-time in a mesh, and without a central server or hardware. “Today, mobile devices are the most powerful computers we own. But you need a different way of building software to enable this,” he explains.
“With many use cases, particularly the military and emerging tech in remote environments that have immediate latency requirements, the concern is loss of data transmission and devices being locked or jammed. Redundancy is critical so if one device is jammed, data can go through another. There must be multiple paths to passing data. The issue of connectivity and syncing data in real-time is a huge problem with profound effects.”
A common cause for a cloud outage is power disruption. While there are many innovations in the works to resolve this issue, microgrids are getting a lot of attention lately. Microgrids help provide clean energy to power facilities, but also improve resiliency by providing reliable backup when traditional means of power are lost or otherwise inaccessible.
“A microgrid is a self-contained electrical network that allows organizations to generate their own electricity on-site,” says Jana Gerber, President, Microgrid North America at Schneider Electric. “Microgrids tie together distributed energy resources [DER] in their system and can operate while connected to the utility grid or in disconnected ‘island’ mode, utilizing software controls and battery energy storage system [BESS] technology to effectively distribute and regulate energy resources.”
“By deploying cleaner, reliable energy sources in the event of an outage, there is significant offset in the carbon emission, which have a health and economic benefit while also reducing the costs to import diesel fuel. The added benefit of this technology is the ability to leverage internet of things [IoT] enabled smart inverters for microgrid control, which allow the microgrids to communicate information to the cloud and optimize the data collection, servicing, and overall management,” Gerber adds.
A common cause for a cloud outage is power disruption. While there are many innovations in the works to resolve this issue, microgrids are getting a lot of attention lately. Microgrids help provide clean energy to power facilities, but also improve resiliency by providing reliable backup when traditional means of power are lost or otherwise inaccessible.
“A microgrid is a self-contained electrical network that allows organizations to generate their own electricity on-site,” says Jana Gerber, President, Microgrid North America at Schneider Electric. “Microgrids tie together distributed energy resources [DER] in their system and can operate while connected to the utility grid or in disconnected ‘island’ mode, utilizing software controls and battery energy storage system [BESS] technology to effectively distribute and regulate energy resources.”
“By deploying cleaner, reliable energy sources in the event of an outage, there is significant offset in the carbon emission, which have a health and economic benefit while also reducing the costs to import diesel fuel. The added benefit of this technology is the ability to leverage internet of things [IoT] enabled smart inverters for microgrid control, which allow the microgrids to communicate information to the cloud and optimize the data collection, servicing, and overall management,” Gerber adds.
-
About the Author(s)
You May Also Like