High Availability Case Studies

MPLS Implementation

As the contract term on their original MPLS implementation was coming to a close, our client, a large financial services enterprise with over 150 sites worldwide, kicked off a worldwide infrastructure consolidation and centralization project and was interested in replacing their global WAN with a more redundant, modular, QoS-enabled solution.

Treadstone Consulting’s engineers were awarded the task of not only design and oversight of the implementation of a worldwide solution, but also of driving and consulting on the RFP process used to determine which carriers would be selected for the final rollout.

After meeting with our client’s staff and assessing their needs from a traffic flow, application, server Infrastructure, and disaster recovery/business continuity standpoint, we were able to present them with a proposed solution that involved using 2 MPLS carriers with separate “clouds” for each geographical region. This would allow not only redundancy and control of the traffic flow, but also allow for each region to have a separate routing table, something our client was very interested in doing, so that a given region’s Internet traffic could flow through data centers in that region. Quality of service would also need support to ensure that the company could implement VoIP and IP Video conferences at a later date. Treadstone’s Systems Architects then crafted the RFP around the requirements set by the overall design and began the process of determining which carriers would be best suited for the job.

After a series of meetings and design discussions, the top two were selected by the client and implementation of project plans were then created.

UNIX Based Trading Application

Our client, an entity at the forefront of the energy trading sector, was in need of a solution to improve the resiliency of their homegrown trading software and the associated web portal, all running on a UNIX platform. Treadstone Consulting’s Systems Architects were tasked with designing a solution to provide security and fault-tolerance at each level of this 3-tier application.

At the front-end layer, Apache-based web servers provide users access to the application. In the 2nd tier, application servers provide access to the database and analyze market data to keep pricing and commodity information up to date. The backend consists of an Oracle database and NetApp storage hardware. As an initial phase Treadstone conducted an assessment of the infrastructure, servers, and components involved in the application and its environment. The focus of this phase was on High Availability and Security for UNIX-based Trading Application of the architecture and interviews with key members of the client’s IT infrastructure and application teams in order to gain a clear understanding of the requirements and dependencies involved in this endeavor. With the assessment activities complete and documented, the Treadstone team began constructing a robust high-availability solution that would meet or exceed all of the requirements as they were laid out in the assessment. The solution that was presented offered redundancy across each tier of the application through a combination of global load-balancing across UNIX server farms at each data center and back-end redundancy through database and storage-level replication.

Maintaining the real-time nature of the market data was accomplished by using multicast streams to make sure that all servers in the farm and users on the floor had the same market data at the same time. Security was addressed by 2 layers of Cisco PIX firewalls at each data center, proxy servers in the DMZ, as well as server load balancer appliances. The complete solution was then presented to the client along with a proposed migration and project plan.

We were then given the task of implementation where Treadstone’s engineers and project management led the charge to put the architected solutions in place.

Telecom Disaster Recovery Assessment

For call center operations and other telecommunications services that are mission critical, Business Continuity Planning is indispensible. It can also be a daunting task, taking valuable resources away from support roles and other projects. Such was the case with one of our clients whose call center is their primary source of revenue. They called upon Treadstone’s telecommunications professionals to assess their needs and their systems and formulate a methodology and implementation plan that would provide protection for this crucial application.

The assessment began by meeting with key personnel within the client’s organization. Then a complete audit of all systems was made. Inventory and architecture documentation was created as a deliverable to the client. This allowed Treadstone’s engineers and the client to gain a full understanding of the existing environment.

With the discovery phase completed, we then began the task of developing a methodology for disaster recovery and redundancy, providing for all conceivable failure scenarios. During this phase, Treadstone also developed a list of infrastructure remediation tasks and additions required to fulfill the requirements of the methodology and created a matrix of these tasks in order of priority and impact.

The final solution was then presented to the client along with all deliverable materials and documentation to form a complete and comprehensive picture of the current and future architecture of the system. After reviewing the deliverables, the methodology document, and the prioritized task matrix, the client was able to make an informed decision about the levels of redundancy desired and the specific impacts of each change. We then estimated the level of effort required to implement the desired solutions and the final proposal was ultimately approved and orchestrated.

VoIP Optimization and SIP Trunk Implementation

As a trading group, our client relies on telecommunications for its life. Before contacting Treadstone, there was an existing VoIP implementation, but the network was not optimized for voice and separate infrastructure was used to compensate.

In addition, standard PRI and long distance T1 lines were landing on Cisco 2621 routers to provide local and long distance service. Also, several T1 lines were used to carry ring down lines to partners and clients. Many of the company’s partners and clients were already connected through a firewalled MPLS connection as an alternative to VPN tunnels. Treadstone was tasked with both optimizing the network and reducing the equipment necessary to operate the VoIP network and with utilizing the MPLS network to reduce the ring down cost.

For this task, we analyzed the existing network, the VoIP solution that was in place, the equipment used, and the circuits involved and developed a solution for the LAN. Also during our assessment, we contacted the MPLS provider and inquired about options for QoS across the cloud, as well as IP-based long distance and ring down options. The provider did have a product we could use in the form of a SIP trunk service. We then put the pieces of the final design together and presented the final deliverable from the assessment to the client. Approval was made and Treadstone began the implementation phase.

For the LAN, older XL-series switches were replaced with newer 2960 switches that support QoS and PoE, enabling the PC and phone to use the same switch port. QoS rules were developed and implemented across the network. The SIP trunk service was implemented on the MPLS WAN by the provider and Treadstone performed the work of integrating it with the Cisco Call Manager. The old 2621 gateways were redeployed as endpoints for IP ring down lines across the WAN.

Wireless LAN for Houston Airport System

The Houston Airport System came to Treadstone Consulting with a need to build a secure, manageable wireless LAN infrastructure in their Administration building as well as in the “disaster room” of a terminal. The scope of the project was immense.

It required supporting the full staff of the Admin building, as well as large numbers of disaster team members in a time of crisis.

The focus of the project was in three main areas. The first primary requirement was centralized management of security and configuration. Manageability is compromised if engineers need to go to each access point or antenna to make changes. The second requirement was for speed and scalability. The solution has to stay reliable and responsive even as the rollout is scaled up to more and more coverage. Lastly, total cost of ownership and return on investment played a major role in the decision to go with the Treadstone solution.

The solution we recommended utilized standards-built, yet innovative equipment from one of our technology partners, along with other key infrastructure components and process controls to provide centralization of site planning, and configuration, ease of use and scalability, security and access controls, and plug and- play installation and automated configuration.

With this solution in place, the Houston Airport System now utilizes the most advanced wireless security available and is able to scale their installation without issues with the infrastructure or manageability problems. It provides them with the ability to perform remote troubleshooting and the integrated firewall provides full security for the LAN while providing the client with the lowest possible cost of ownership for an enterprise wireless solution.