> KEMP Bare Metal
KEMP Bare Metal
- Cisco UCS Series
KEMP Technologies is a leader in the Application Delivery Controller space because of a highly valuable Application Delivery solution suite catered to meet the complex technical business needs of its clients. Cisco is a visionary in both the networking and blade computing arenas and has, since its beginnings, been instilled with the concept of designing solutions of unprecedented value driven to address pointed and specific customer challenges.
Application delivery is often described as combining specialized technologies in order to provide high availability, security, visibility and acceleration. This concept involves the optimization of application traffic along with the ability to deterministically make the best decisions on how that traffic is handled. An intelligent Application Delivery Controller plays an integral part in making this possible. This, combined with a framework designed for high-speed and low-latency communication provides an optimal environment for Application Delivery. The following discusses the innovative solution, LoadMaster for Cisco UCS, that KEMP Technologies and Cisco have partnered to deliver to their valued Large Enterprise customers.
With the advent of cloud computing a number of new technologies have emerged that veritably re-painted the landscape of enterprise core IT infrastructure. Not the least among this ilk is blade computing and Cisco’s innovative contributions through their Unified Computing System (UCS) platform.
The Role of Blade Computing in the Modern Large Enterprise
Blade computing is a methodology that leverages an integrated architecture encapsulating power, cooling, cabling, networking, storage and consoles. This approach spreads the costs of these components across several servers in an enclosure. This has the following benefits:
- Reduction of overall cost per server
- Reduction of number of parts required per server
- Increased server density
Cisco takes this technology several leaps and bounds forward with their innovative UCS platform and accompanying unified fabric. This allows for the consolidation of IP, Storage and Management networks using a “wire-once” philosophy. The resource allocation for things such as bandwidth and I/O configurations are then administered dynamically via the system’s integrated administration interface. This allows for all servers to host the same workloads by simply adjusting configurations via embedded software. The approach of leveraging a single network technology for IP, Storage and Management not only reduces acquisition CAPEX and TCO but also dramatically increases performance by eliminating reliance on external network architectures as well as reducing the number of interfaces, cables and switches required to support the infrastructure.
The Role and Challenge of the Application Delivery Controller
One of the driving forces behind the selection of this type of sophisticated platform for the modern Large Enterprise is the need to deliver scalable business critical applications in a high performance and highly available manner while reducing total cost of ownership. To this end, a supporting key requirement is the ability to manage and optimize user traffic across the infrastructure. This has historically been the job of a purpose-built Application Delivery Controller (ADC) appliance. This appliance will generally consist of fixed-configuration hardware – combining commodity X86-based and proprietary ASIC-based components with customized software to create the functionality framework of an Application Delivery Controller.
This approach fills the aforementioned needs and provides stable performance but has a relative lack of scalability and “mobility” that is synonymous with the consolidated infrastructure that it provides front-end services for. This can create challenges for the applications that reside on the infrastructure as well as server administrators and effectually negate the benefits realized by blade computing being delivered by a platform such as UCS.
Application Delivery Controller Virtualization
To address the complex challenge, KEMP Technologies pioneered the utilization of hypervisor abstraction in the ADC context in order to facilitate “virtualized” ADCs, making the aforementioned “elasticity” possible in this ecosphere. This abstraction essentially divides resources down to the least common denominator of compatibility and allows for separation of the operating system from the physical hardware. While this brings many benefits, the drawback is that the actual abstraction, which is a de facto mechanism of virtualization, inherently “blocks” the ADC from fully leveraging the foundation hardware and networking performance optimizations in the underlining infrastructure that are engineered to improve overall performance and scalability. Additionally, the innate “flexible” nature of hypervisor platforms (e.g., VMWare, Hyper-V, KVM, Oracle VM, etc…) can actually limit physical server resource availability for the ADC which can result in bottlenecks and overall performance degradation.
LoadMaster for Cisco UCS
To surmount this challenge, KEMP has developed a specialized version of its renowned ADC Operating system (LoadMaster for Cisco UCS) that is optimized to operate natively on any Cisco UCS server! This optimization allows KEMP to leverage the high performance hardware and network “transport” capabilities that are already a part of the unique UCS fabric in order to provide stratospheric ADC performance levels that are without match in the industry. Due to the fact that the underlining hardware platform is truly dedicated to the ADC tasks, there is no need for abstraction and stable performance service levels can be “guaranteed” for various metrics (e.g., throughput, SSL TPS, connection concurrency, etc…) As the only Cisco Developer Network (CDN) Partner in the ADC space that has been vetted by Cisco to be IVT (Interoperability Verification Testing) approved, painstaking effort has went into the development of this solution to ensure that it provides Large Enterprise customers with the best possible experience. As can be seen in Figure 1.1, which depicts an integrated Cisco UCS infrastructure, the communictaion within the UCS fabric operates at ultra-low latency delivering ultra-high performance levels.
On the other hand, the infrastructure external to the Cisco UCS environment is subject to a number of variables that contribute to and create congestion and latency. When the Application Delivery Controller is deployed at the network layer in an environment leveraging the Cisco UCS platform, it too suffers from the aforementioned negative variables, contributing to the bottle-neck for overall performance that the network layer shares responsibility for as a whole. By leveraging KEMP LoadMaster for Cisco UCS, which integrates the ADC function directly into the high-performing UCS fabric, many benefits are made possible.