The Fabric Automation software now comes embedded in two newly announced switches, and has been integrated into VMware, Microsoft and OpenStack orchestration software. This greatly reduces provisioning time, and improves the efficiency of operations.
Extreme Networks has announced upgrades to their Extreme Fabric Automation software, which they believe significantly ups its value as a differentiated solution in the market. They have announced two new high-performance switches – the SLX 9150 leaf switch and the SLX 9250 spine switch – which are the first to have the software ship embedded in the switch. They have also announced new integrations with VMware, Microsoft and OpenStack orchestration software. The new product and integrations both improve provisioning time and operational efficiency.
Extreme Networks introduced their Fabric Automation a year ago, and enhanced it late last summer with tenant provisioning.
“Fabric Automation is like middleware,” said Dan DeBacker, Director of Product Management at Extreme Networks. “It’s a microservices architecture, which acts as a full lifecycle management type tool, becoming the operations procedure for the fabric that provisions the underlay and overlay. You can access it with CLI, but in most cases, it’s being integrated with APIs. We are now taking it to the next step with the 2.1.0 version.”
One of the keys here is the two new high-performance switches embedded with guest VMs – the SLX 9150 leaf switch and the SLX 9250 spine switch. These are not simply new products being released with the software. The embedding of the software within the new switches makes them a fundamental part of the fabric, and helps make it even faster.
“The cool thing is that the fabric automation is embedded into these spine and leaf switches,” DeBacker said. “With our Extreme Insight Architecture, they now become part of the IT fabric. Three simple commands and the fabric is fully provisioned. You never have to do any configurations. Earlier this summer we had a customer deploy 28 switches in a 5 stage Clos [network], which took 161 seconds, fully provisioned. The most important piece is that it was provisioned with zero errors. It validates the physical typology, and if there is an error, it tells you. It provisions everything on the underlay and overlay, so you can then do tenant provisioning. That takes 15 seconds, in three commands.”
DeBacker indicated that making the two new switches part of the automation process doesn’t make all their previous switches obsolete.
“The technology ships with the switch, but it is not coupled with the actual switch OS,” he said. “We update the Fabric Automation every 12 weeks. We won’t ask customers to update every 12 weeks. This means that if a customer doesn’t have these new switches, they can still use the technology on other switches. They can host it externally. We won’t force people to do things a certain way. We design products to fit their business, not our business.”
DeBakker also emphasized that the Fabric Automation is not a product.
“It’s not something that the customer has to purchase,” he said. “We consider it table stakes, using automation to deploy your fabric. We provide this automation tool as a feature. You don’t have to install it on a server. Since it’s a feature and not a product and you don’t have to pay for it, it becomes a natural part of the solution, and you have the flexibility around how you want to use it.”
The Fabric Automation’s value isn’t limited just to provisioning. DeBacker noted that while the new SLX spine and leaf switches are built on Broadcom’s Trident 3 switch ASICs, following the announcement a month ago that Broadcom would be Extreme’s preferred provider for enterprise campus networking solutions, what matters isn’t the hardware, but how Extreme uses it leveraging their software.
“We aren’t here to differentiate on hardware,” he said. “It’s not just about the spine and leaf hardware. It’s what you can do with it. So this bringing it into the automation is a big step. This isn’t just for provisioning, but for longer term operations. On top of the automation we are differentiated by our Insight Architecture, a dedicated data path from the ASIC into the guest VM. Our whole idea here is how to reduce the ‘mean time to innocence’” – a tongue in cheek term that demonstrates that the network admin isn’t to blame if a problem develops during operations.
DeBacker also highlighted the importance of the newly announced integrations between the Fabric Automation and VMware VCenter, Microsoft Systems Center and OpenStack orchestration software.
“VMware is logical because they are the overwhelming market leader,” he said. “Microsoft is strong in some verticals where we are also very strong, like education. Service providers use OpenStack, so that makes sense. Providing integrations for them all gives us integration points across all these different markets.”
Each of these integrations is a separate microservice and leverages the application’s fabric awareness. Additional integrations will be available in future releases.
For Extreme’s channel partners, the new capabilities in the Fabric Automation software provide further differentiated value and a way to avoid the trap of being forced to compete by cutting prices.
“Partners want a differentiated story they can tell the customer, and which they can demonstrate easily to show value,” DeBacker said. “Our competitors all have good 25 and 100 GB switches. Partners don’t want to compete as a commodity thing and discount. They want to show a value that is differentiated. They can show the customer that they can provision their entire fabric in less than five minutes. That’s the biggest message, that this is something that is different, and that the products are not all the same.”