Kaloom’s Container Networking Dramatically Improves Overall Application Performance

Kaloom announces integration of OpenShift and Kubernetes, delivering major CNI and container networking improvements

Kaloom, an emerging leader in the automated data center networking software market, announced the complete integration of OpenShift and Kubernetes into the Kaloom Software Defined Fabric (SDF). At KubeCon, Kaloom is showcasing its container networking improvements and enhanced version of the Kubernetes CNI, named Kactus, to better sustain mission-critical applications, high-availability and support emerging container networking functions.

Data is growing faster than compute and much of that growth is from computer-to-computer communications supporting mission-critical applications. The network must efficiently deliver high throughput or it could become a major bottleneck, making it the most critical piece for improving data center application performance.

Kaloom recently announced the availability of a fully automated programmable data center networking fabric for enterprises, cloud providers, gaming companies, data center operators and 5G wireless providers. The Kaloom SDF is fully containerized which contributes to the scalability and portability of its solution. Kaloom SDF is integrated into both OpenStack and OpenShift/Kubernetes for a unified virtual machine and container operating environment, simplifying the application deployment in data centers. It allows data center providers to host any virtual machine or container-based application and improves overall application performance, especially for container-based apps.

Read More: FreakOut Announces Expansion, Appoints New Leadership Team in Russia & CIS, Middle East and Australasia

High availability Kubernetes applications must be capable of providing 24/7 availability over long periods of time. In this operational environment, changes in the data center network are considered normal occurrences whereby the provisioning of additional NICs or ports on application-servers, or their deletion, can have adverse effects and cause link or port failures to occur. Consequently, the Kubernetes CNI must be capable of dynamically discovering and adapting to changes in the networking environment. The current CNI only permits the discovery of the network topology at start time. In the event of a planned or un-planned change to the network, the Kubernetes application needs to be shut-down and restarted, thus introducing significant disruption to the application.

Typically, in Kubernetes each pod only has one network interface other than the loopback. This is perhaps sufficient for basic control plane functions using a request-reply paradigm. However, several applications (i.e., Firewalls, Video Transcoding) require the use of multiple network interfaces per pod, whereby packets may enter via one interface and exit through a different one.

Kaloom’s integration into Kubernetes and OpenShift is via its development of an enhanced multi-network Container Network Interface (CNI) plugin called Kactus which solves several limitations in Kubernetes. It enables the attachment of multiple network interfaces to pods in Kubernetes, as well as dynamically discovering and adapting to changes in the networking environment.

Kactus is also integrated with both Open vSwitch (OVS) and the Kaloom Virtual Switch (KVS), which is an embedded accelerated virtual switch that provides performance improvements in terms of throughput and latency when compared to OVS-DPDK.

Read More: Five9 Customer Service Index 2018 Reveals a Rise in Customer Expectations and Willingness to Share Data

Kaloom SDF collapses the data center POD into a single, or a 1-tier, topology to enable better scaling for SDN networking and management. Kaloom SDF is SDN-controller agnostic and is integrated into the OpenDaylight SDN controller, making it well suited for an SDN style of management that alternative container networking solutions struggle with. Kaloom provides the best of breed with a centralized SDN control and peer-to-peer data plane (hybrid model), giving excellent traffic engineering capabilities. In addition, it provides better availability and scale notably versus host-based BGP solutions, tenant isolation and a programmable data plane. Advanced services chaining can be offloaded into the data plane for better overall application and network performance.

“Kaloom’s engineers work diligently to improve container networking and we believe the data center networking market will be excited to see the improvements in terms of scale, availability and manageability. Furthermore, Kaloom is committed to work with the open source community and container networking leaders to contribute innovations upstream that will benefit partners and customers,” said Suresh Krishnan, chief technology officer at Kaloom. “Our real-time in-band telemetry solution can be used to dramatically improve the network visibility of these container environments.”

“Ongoing developments in cloud-native applications give every indication they will continue introducing demanding requirements to be met by the networking and computing infrastructures that support them. They will require increased versatility in the networking environments they integrate with, and their networks will need to become even more resilient and robust as new workloads are introduced,” said Paul Parker-Johnson, Principal Analyst with ACG Research. “Kaloom has seen these requirements coming and incorporated important innovations addressing them in its Kactus plug-in for the Kubernetes CNI. Enhancements include improved performance, resiliency, scaling and integration into heterogeneous networking environments that will be important in both computing and networking infrastructures moving forward.”

Read More: Executive Awareness: The Key to Boost AI Adoption In Any Industry