Our platform has developed since 2000 according to our customer needs and available technologies. With Microsoft as core technology, we moved to .NET 2.0 architecture based on services and AJAX technology as it became available as this proved to be an experience rich and fast environment to execute the clock auction methodology. A central auction engine coordinates auctions in real time in order for anyone to participate anywhere in the world in any language. Plug-ins per auction type are available so we can develop custom made auctions if needed. Essential elements of the design were:
Scalability: fast growing number of auctions, users and auction types must be possible
Flexibility: different auction types must be pluggable to stay ahead of market development
Stability and safety: Auctions should not crash or be hacked. Never
Traceability: All actions of all users should leave a trace in the system for review or reporting
Hardware
The Negometrix system runs on trusted and proven hardware vendor like HP. Each server has built-in hardware redundancy – redundant power supplies, redundant chassis fans, hardware RAID adapters. Also all redundant parts are “hot swappable” – in the event of failure they can be replaced without stopping the servers which means that there is no downtime for the customers’ using the system, and most important no downtime during auctions. The hardware RAID adapters provide fault tolerance for the hard drivers that the servers use, ensuring that if a hard drive fails, there will be no data lost and the server and applications will continue to function normally. The hardware RAID adapter also offloads the tasks of providing fault tolerance from the operating system thus increasing the servers’ reliability and performance.
Further, the Service Level Agreements (SLAs) for those servers guarantee an onsite replacement of failed parts (RAM modules, processors, main boards, hard drives, fans, etc.). The replacement should occur from 4 hours up to 1 day since a part has failed. That way we can ensure that in a worst-case scenario, there will be no more than 1 day of downtime for the services, applications, and sites running on the failed server. But thanks to the proven hardware we use, there has never been such a situation.
Operating Systems, Software, Updates
As a Gold Certified Microsoft Partner, our technology partners at BSH –responsible for the daily maintenance of the servers use exclusively Microsoft server products, and applications to provide Negometrix customers a reliable and secure environment with superior performance.
All of the servers are regularly updated with the latest critical, security, performance, and application-specific updates. This makes us sure that the platform we provide to our customers is up to date, secure and reliable. Before applying the updates to the production servers they are extensively tested in our test lab to ensure that there will be no problems or conflicts with the currently running applications.
BSH is currently in the process of migration of all the servers to Microsoft Windows Server 2008. That is the latest, most secure, most reliable, and best performing operating system from Microsoft.
Monitoring
There is no such thing as a perfect software or unbreakable system. Errors happen, software fails, systems go down. What we can do is to keep an eye on all the components running and prevent failures before they happen. BSH is pro-actively monitoring its servers, applications, services, and web sites for a wide range of indicators – from servers’ fans speed up to Web sites’ performance. When some component does not match the baseline of its normal operation, an alert is triggered, and our support team reacts to this alert, by finding the source of the potential problem and solving the problem. That way we know about the problems and prevent them before they have occurred, thus ensuring there is no downtime for the applications running on our servers.
Backup and disaster recovery
Backups are an integral part of each successfully running infrastructure. Having a good backup of your data in a secure location, you can survive and recover from almost any disaster. Using Microsoft System Center Data Protection Manager 2007, BSH ensures that all the Tenders in the Negometrix system are backed up and can be recovered in the event of failure, data lost or other kind of disaster. Backups are stored on two separate places. The short-term protection data is stored in our collocation on fast hard drives, providing extremely fast method for recovering. The long-term protection data is written to tapes and kept offsite in a secure location, ensuring that there is a solid backup over the time.
Internet Connection and Power Supply
The internet connection and power supply for the servers are provided by our hosting partner – Orbitel, which is part of Deutsche Telecom. The servers for our customers are connected to the Internet using redundant routes, so that in a case of failure of one route, the connection automatically start to flow through one of the other available backup routes.
In the case of a power outage, there is a UPS network to support all of technical equipment. If the outage lasts longer, there are diesel power generators, which can support the load.
Security
Security has many aspects. Physical access to our server machines have only authorized personnel from BSH which first passes a check in our collocation partner before been granted access to the servers. Remote access for management, maintenance, software updates and other support activities is also granted only to authorized BSH personnel. All of this activities are performed over a secured and encrypted connection. All access to our servers for support and other similar tasks is logged so that we have history of who did what and when.