The first, true Linux OS for data center networking
Cumulus Networks is transforming networking with the industry’s first, full-featured Linux operating system for networking hardware. As application requirements continue to grow, businesses are looking for ways to scale data center operations, including more efficient automation and monitoring of agile, resilient and fast networks in a cost-efficient manner.
Zstor Software Defined networking
Solution: SwitchES with Cumulus Linux
Cumulus Linux is a software-only solution that enables users to take full advantage of the latest, industry-standard networking hardware while enabling the latest Linux applications and automation tools. Cumulus Linux reduces complexity and simplifies network management while delivering new levels of scalability and flexibility to the data center.
How is it different from traditional network solutions? The new network paradigm is an open ecosystem where businesses can choose between various industry standard networking gear, network operating systems and network applications to meet their business needs.
Cumulus Linux fits in the middle between networking hardware and applications. Cumulus Linux is different from other network Operating Systems because it is not a Linux-based OS, it IS Linux and offers the entirety of the Linux experience on networking hardware. The control and management planes are native Linux while the data path is accelerated using switching silicon. All Linux applications and device abstractions are maintained. This allows existing and custom Linux applications to run natively on the switch. In addition, new applications can be developed solving customer challenges and enabling further open source innovation.
Cumulus Linux enables the following:
Latest datacenter networking architectures
Leverages Linux standard Layer 2 and Layer 3 control protocols, enabling the latest leaf/spine and network virtualization architectures while interoperating with traditional architecture
Lower Total Cost of Ownership (TCO)
Fits in an open ecosystem where customers can choose from a broad set of bare metal hardware and applications, enabling higher capacity and availability at a lower cost
Large scale orchestration and automation
Leverages Linux advanced management, scripting, and automation tools such as Chef, Puppet, CFEngine
Customers can now choose from a large set of Linux monitoring tools such as Collectd, Ganglia, and traditional SNMP tools. Beyond Linux native applications, Cumulus Prescriptive Topology module ensures physical and logical topology consistency
New wave of innovation
Business can write their own applications to run on our switches, bringing about new sets of applications through collaboration
Cumulus Linux is the missing piece needed to build a modern, scalable data center, bringing the same paradigms of manageability, clustering, monitoring and orchestration that are applied at the server and virtual machine layer today to the network, and realizing the full promise of the Software Defined Data Center.
Key Cumulus Linux Functionality
Linux Operating System
Resilient Linux Modular Operating System with fault containment, extensibility, server-style upgrade/ patching, server-style process restart/termination.
Linux extensibility - Any language supported in Linux today, includes scripting with Bash, Python, Perl, Ruby.
Switch Hardware Abstraction Layer
Hardware acceleration including IPv4/v6 routing tables, ARP tables, Port/VLAN information, and bridge tables, resource allocation/management.
Management and Automation
Management tools such as RADIUS, TACACS+, LDAP, DHCP, v4/v6 DHCP relays, NTP, Zero-touch install using ONIE loaded on bare metal switches, zero-touch provisioning.
Third party tools can be added for advanced automation and troubleshooting including Puppet, Chef, CFEngine.
Monitoring and Troubleshooting
Traditional monitoring with SNMPv2/v3 and network specific MIB, Analytics with SPAN, ERSPAN, ACL-based counters, real time queue-depth and buffer utilization reporting.
Ease of use with Prescriptive Topology Module, taking a user specified network graph as a basis to ensure topology consistency, proper cabling termination, and neighbor/prefix reachability.
Third party tools can be added for advanced monitoring including Collectd, Ganglia, NetSNMP, Nagios.
Routing and Switching
Enhanced Quagga IPv4/v6 routing suite including OSPFv2, OSPFv3, BGPv4/v6.
Bridge management including STP (IEEE 802.1q), RSTP (IEEE 802.1w), Bridge Assurance, BPDU Guard, IPV6 Router Advertisement Guard, VLAN trunks (IEEE 802.1q), LACP (IEEE 802.3ad), LLDP, IPv6 neighbor discovery, unicast/broadcast storm control.
Access Control Lists (ACL) L2-L4 classification, CPU protection through hardware enforced ACL-based rate limiting.
Class of Service (CoS - IEEE 802.1p), CoS based queuing, CoS based scheduling (DWRR, WRR, Strict Priority), ingress ACL based classification/policing, shared buffers, bandwidth allocation per port group.