Expertise for where network communications is going

What sets Northforge Innovations apart from the rest is our Intellectual Capital.

It’s how we look into future of networking and how we develop our expertise today to meet the needs for tomorrow’s networks. Instead of only being focused on today’s requirements from our customers, we ask “what expertise will be needed for networks in the next few years.”  Our customers need to be ready to implement the latest technology advances and we will be ready with the expertise to help them with future challenges and their most complex issues.

When cloud computing was in its infancy, Northforge was already looking ahead and getting ready so that we will be ahead of that technology curve. We saw this cloud computing movement as a potential game changer. We searched to find the right experts and give them the experience they needed to be leaders in this new space.  Now with everyone moving to the direction of cloud computing, Northforge is already there, leading the way and helping its customers to go in that direction.

Another component of our Intellectual Capital is our approach to problem solving through increased performance. We aren’t satisfied with handling the simple problem. We excel when we are given a complex challenge, the kind of challenge that most companies shy away from. Not Northforge since it’s that tough challenge that gets our passion to advancing network communications racing. The challenge isn’t coding a basic application. Instead, give us the complex challenge of when you have a wireless network that’s dropping calls or when you need to increase performance 10x or whatever difficult performance challenge that you are facing.

Intellectual Capital is a passion that’s inside each of our Northforge employees. Here are some examples on how we pushed the limits from our employees:

 C.C., Software Designer

Monitoring System Technology

“When a customer came to Northforge needing to enable the use of one network services processor (NSP) in sharing mode between two or many NSPs, affecting the device lookups and configuration transactions with minimal delay, they also came with high expectations. Besides satisfying the customer’s need, we wanted to provide a solution that exceeded their expectations, but the feature didn’t leave much margin for exceeding their already high goals.

The feature to be developed required special knowledge and experience on the existing system and some creativity to bypass some physical and logical constraints that were preventing the access to the NSP from different NSPs. Our main challenge was to let more than one NSP configure the NSP, since there was only one physical channel to configure it which can be connected to only one NSP. Our solution was to use the physically connected NSP as a configuration collector/dispatcher for the other NSP.

I used my prior experience on the system to create a better fitting solution in the existing environment, including UML to design the solution, Cavium OCTEON processor, C/C++, plus I developed expertise on the OCTEON NSP. As a result, all of the NSPs involved were able to access the NSP, sharing it, with no measurable delay in the NSP lookups. We released the solution to the customer within their time requirement, even after we uncovered unforeseen obstacles.

With some creativity, passion, commitment, hard work and technical skills in UML, Cavium OCTEON processor and C/C++, we were able to deliver significant results which exceeded the customer’s already high expectations.”

J.S., Senior Software Engineer

Network Security Technology

“Our customer, a leading networking component provider, asked us to make design modifications to support OpenSSL DTLS (Datagram Transport Layer Security), so that UDP traffic processed by their hardware is as secure as TCP traffic. The project would be validated against their baseline test suite, including the performance benchmark tests, and one of their goals was to at least match the measured performance of the previous product baseline.

The project had a host of challenges, but I applied my expertise and knowledge in network security to fully meet the project requirements. Using my Ph.D. in Computer Science and 20 years of telecom experience in network security and cryptography as a foundation, I applied my deep understanding in Public-Key Infrastructure (PKI) — a key factor in the OpenSSL design principle, and developed several hundred test cases to verify security features.

To solve the challenges, we added new features to provide the state-of-the art OpenSSL DTLS services to enable secured UDP traffic handling for their multi-core networking processors; enabled TCP and UDP dual-stack functionalities with backward support of full TLS services; enhanced their PKI system for extended timeline coverage and increased security strength and cipher algorithms; and completed the performance sampling on all metrics, improved TCP/IP traffic throughput performance as compared with the previous product baselines. Average DTLS throughput for this project increased nearly 100 Mbps.

I was determined to exceed the customer’s expectations in all categories — software feature scope, software quality (zero-defect), and time-to-delivery – and we exceeded them all.”

F.L., Software Developer

Secure service sharing

“A customer was moving to newer OCTEON chips from Cavium to produce faster and more efficient version of their best in market consumer product, and we applied our vast knowledge in Wi-Fi open source, Linux kernel, access points and our Cavium experience to allow them to release a newer version of their OS on the newer and faster OCTEON chip with the Linux kernel. We improved throughput with a newer kernel (2Gbps+ Wi-Fi traffic) varying with hardware configurations); provided newer and improved features in the kernel and newer versions of open source application used e.g. IPTables et al; and significantly improved upgrade and downgrade of firmware version across various supported platforms.

But it wasn’t without its challenges. Such as a large prewritten code base (Kernel + drivers + applications) with an aggressive timeline to debug through code manually and the challenge to reproduce scalability tests which required large volumes of network traffic. So we used our experience of Linux on OCTEON, our open source knowledge and system build, and our networking protocols understanding to support and debug their new release of their firmware which mainly migrates to Cavium’s new chip and SDK (Linux kernel 2.6 to 3.10).

We fixed issues which arose with 3.10 kernel so that new software would work appropriately with the customer’s existing features, ranging from issues in Atheros Ethernet driver to those in open source Linux applications such as Netfilter, IPTables etc. as well as in the Linux kernel itself. We also tested the product for extended periods of time for stability with loads of traffic.

As a result, we stabilized the ported code base to a much mature level than when we started, making many features of the existing firmware work in the newer code base.”