5701 Lonetree Blvd.
Rocklin, CA 95765
|
(916) 920-4730 |
Having a core set of high-performance servers to deal with routine traffic and a dynamically provisioning cloud platform for seasonal and spike traffic is the ideal infrastructure for e-commerce and serious Web app companies.
Our production network consists of dozens of custom built brand name servers, network switches, storage devices, security hardware, etc. We maintain a PCI data security standards for all our e-commerce customers and provide HIPPA compliant security facility for our medical service customers.
We have been in a seven-year business relationship with one of nation’s top-10 data centers with the leading IP4 peering and nationally recognized IP6 leadership. We own and operate all our equipment in the same Silicon Valley location where companies like CISCO host their global infrastructure.
We have long-term partnerships with both our customers and vendors such as our data center provider.
We use Barracuda Network’s high performance, DIRECT PATH ROUTING enabled load balancers with stateful failover capabilities.
We use IBM rack-mount servers with 32GB to 128GB RAM, quad-core processors, high-throughput Gbit network interfaces and 15,000 RPM SAS drive arrays with remote power and KVM capabilities. We are always exploring new hardware to bring in performance at different levels of the complete solution. For example, we are now exploring the new Linux kernel feature bcache that works with Solid State Drives (SSD) to increase disk subsystem performance. Bcache is a filesystem agnostic patch for the Linux kernel to use SSD drives to cache other block devices such as a raid array. By default it won’t cache sequential I/O, just the random R/W that SSD drives excel at. By keeping up with both hardware and software advancements, we are able to create very performance tuned setup for our customers.
Using a state-of-the-art cloud infrastructure we can and have dealt with large-scale DDOS attacks against our customers by black hat SEO thugs who deploy distributed denial of services to attempt to take down competitors who have organically superior position in search results.
Using backup appliances in multiple geographically diverse locations, we back up the Web apps, database contents incrementally on a daily basis and complete weekly backups are made automatically to ensure disaster recovery.
A critical part of successful development and testing is to have mirror infrastructure for development and testing.
Unfortunately, it is ridiculously expensive to maintain a second production network. So we use a virtual machine (VM) based development and testing facility in-house to simulate similar environments.
We mirror live data in our development network so defect (bug) fixing becomes very realistic exercise as data-driven bugs are caught easily due to the exact up-to-the-second nature of our data replication. At the same time our development efforts have ZERO effect on production environment.
We have automatic nightly builds from a git repository of development branches to showcase what is working, broken or coming up the pipe-line. It’s a beautiful and must-have thing to have when you have dozens of software engineers building your apps.