I recently had the pleasure of catching up with the team at Big Switch networks, where I had was run through the product base. I am going to cover off three areas here around the organisation, with three main areas:
- Who are Big Switch?
- What business problems can they help with?
- How does it work?
Who are Big Switch?
Big Switch Networks was founded in 2010 by Guido Appenzeller and Kyle Forster, with around 200 employees today. To date they have focused on equity funding with around $120 Million raised in 5 rounds (from http://www.crunchbase.com). The company currently has revenue growth of 115% YoY.
Big Switch has two main products being, Big Monitoring Fabric and Big Switch Cloud Fabric. They have focused on software defined networking. This is where rather than differentiating the platform based on the hardware or custom ASIC’s, the focus turns to utilising commodity hardware and differentiate in the software. This in turn allows greater advances in features and allows the running of software on many different server providers.
Using bare metal servers and open networking is in line with how hyper-scaler’s are managing their networks, utilising a centralised control plane. This where the company takes its name from, using a centralised management creates one “Big Switch”.
(From the Big Switch web site: http://go.bigswitch.com/rs/974-WXR-561/images/0657-27FA2_BigSwitch_BCF-PoV_WP_WEB.pdf)
- Digital Transformation – the ability to support digital transformation clearly relies on the automation of all steps in the process. By creating a software driven system that has an API for integration and automation, a software defined network supports this business priority.
- Increased Operational Flexibility – Automation allows business processes around change management to be built into workflow automation tools, also creating more confidence in changes.
- Reducing Capital Expenditure – By using commodity hardware the capital expenditure is much lower to enable a software defined platform, the customer can also use their existing server provider to deliver the capability. Also there is no longer a requirement to purchase dedicated routers and switches.
An overview of the capital expenditure benefits are below:
How does it work?
Big Switch have focused on two technical platforms to support their customers:
- Big Cloud Fabric
- Big Monitoring Fabric
The Big Cloud Fabric is the switching component of the platform, the SDN underlay in the platform and is not a competitor to NSX. It can be used for a network underlay using Dell physical switches or HPE switches. The software runs on commodity hardware with an x86 controller, with a Clos (after Charles Clos) architecture. The platform is straight forward to setup:
- Deploy controller
- Provide MAC addresses of network devices
- Download light weight Network Operating System
- This is then managed as one large Logical Switch (Big Switch)
- Integration is via API call with support for vCenter integration
- Scales to 128 Leaf switches, 64 racks, large management domain
- Hyper-scale inspiration is to prevent fault domains
- VXLAN is used within the pods
Traditional vs Big Switch design is shown below:
The Big Monitoring Fabric is where they compete with Gigamon, or Netscaler. This supports building out of band monitoring of a production network, compelling story and secured a foothold in various accounts.
The big monitoring fabric is a next-generation network packet broker (NPB) that leverages SDN principles, Open Networking switches and a high-performance x86-based DPDK service node to provide feature-rich, scale-out data center monitoring at up to 50% less cost than traditional NPBs. Big Monitoring Fabric supports 1G, 10G, 40G, 100G for the most demanding and high volume network monitoring and security environments. Customer use cases for Big Monitoring Fabric include:
- Monitor every rack
- Monitor every location
- Monitor mobile/LTE networks
- DMZ/Extranet Inline security.
The platform has a comprehensive alerting and monitoring platform with strong visualisations that the customers need to view and manage their environment:
Other items covered
We also discussed Openstack as part of the conversation and it is interesting to see that adoption of Openstack is higher for Service Providers. China is a huge adopter of Openstack. 1/3 of attendees at the Openstack conference were from China, this could be due to lack of VMware adoption (my hypothesis).
What about container based environments? Going to DockerCon and Openstack events, conversations are about containers. In container environments BigSwitch owns the physical and virtual switch. In VMware environments BigSwitch only owns the physical switch and VMware with NSX is predominantly used. With this setup a common architecture is able to provide Enterprise container networking. This is used in Chinese car manufacturers, and a lot of R&D in Silicon Valley. The platform is certified to work with Mesos, Openshift and Docker.
I finished with a question around how Big Switch recommends customers adopt automation and orchestration in the enterprise network.
The biggest companies all wished they had started earlier, benefits are so good and clear and keep coming
Thats it, if you want to try out the platform head over the online lab and sign up, its free to try out and an excellent resource, http://labs.bigswitch.com
Overall the platform is making inroads into enterprise customers, and the centralised monitoring platform, will help to make the observations across the enterprise wan easier. With an approach that uses low cost hardware, customers that are not using SDN yet should strongly consider this as an option.