ALTOROS CUTS LARGE-SCALE DATA PROCESSING TIME FROM WEEKS TO HOURS
Using Cloud Foundry platform as a service, Altoros helps organizations build next generation application delivery capabilities, empowering DevOps, developers and data scientists who extract hidden value from terabytes of information within organizations. The most common business problems solved by Altoros are high cost and slow “idea to production” cycles for new software development across the traditional line-of-business applications, and next generation, dataheavy applications that uncover patterns, predicting outcomes, assessing risks, detecting fraud, or building selflearning recommendation engines.
As implementation partner of leading Hdoop, NoSQL & NewSQL providers, including Cloudera, Pivotal, Hortonworks, Basho, Couchbase, Datastax, and NuoDB, Altoros implements and scales next generation, PaaS-driven applications for data mining, analysis, and visualization. Altoros serves a diverse range of customers. For one of its customers, a provider of a cloud-based image recognition platform, Altoros developed a next generation image recognition platform that could offer 5x to 10x improvements when it comes to the cost of processing any given image, especially during spikes of demand.
One of the early beta users of this next generation platform was The Wake Open, a national wakeboarding championship. Its interactive mobile and web applications leverage the newly designed image recognition platform. Due to promotional campaigns on ABC TV Sports during the 2013 season, usage patterns were highly un-even throughout the season. To ultimately reduce the cost of image recognition transactions, Altoros required a Cloud Computing IaaS partner with a high-performance back-end network and granular configuration options for each of the compute, storage and network building blocks of the cloud computing services.
We were impressed with the business impact of ProfitBricks’s granular configuration options for each of the compute, storage and network components in their IaaS platform.
The Altoros team met ProfitBricks at Interop 2013, the leading infrastructure event. They were impressed with the ProfitBricks technology infrastructure and began evaluating customer projects to find those that may benefit the most from unique capabilities of the ProfitBricks platform.
Altoros introduced ProfitBricks to its customer. Ninety days later, the next generation image recognition platform was deployed, and ready to accept workloads. The Wake Open applications enjoyed a great success – attracting more than 3 million unique visitors, many of whom used the image recognition system utilizing ProfitBricks infrastructure, over the website’s 3-day lifespan. To ultimately reduce the cost of image recognition transactions, Altoros required a Cloud Computing IaaS partner with a high-performance back-end network and granular configuration options for each of the compute, storage and network building blocks of the cloud computing services.
Altoros expanded its relationship with ProfitBricks to take advantage of its elastic and high performance computing features for many of their other clients.
The Server We Need First-generation IaaS providers force customers to pick from a short list of instance sizes and network configuration options or packages. As a Cloud Computing 2.0 provider, ProfitBricks offers customizable servers and networks. You can select the desired amount of CPU cores, RAM and storage required for your workloads and deploy it on a network designed by you that scales to your exact specifications.
We Live in a Big Data World – Performance Matters
Processing Hadoop and Big Data queries on firstgeneration IaaS providers takes significantly longer to process due to their use of Ethernet as a network transport protocol. The performance problem with Ethernet comes in its basic structure. With hierarchies of subnets connected by routers, network packets have exactly one path they can traverse between any two points on the network. While you can slightly increase the size of the pipe between those two points, you still, fundamentally, have a single path. „Performance of the underlying compute, storage and network is mission-critical for data-centric applications that leverage Cloud Foundry PaaS, and scale-out NoSQL/ Hadoop data stores.“ A typical Ethernet-based network runs at 1 Gigabit per second (Gb/s), and a fast one runs at 10 Gb/s. A dual-channel InfiniBand network, like the one that ProfitBricks has deployed runs at 80 Gb/s making it the perfect platform for Big Data and other performance intensive clusters.
Incredible Scalability and Flexibility, Including Live Vertical Scaling
Live Vertical Scaling is the ability to increase the capacity of existing hardware by adding resources - for example, adding CPU cores and RAM to a server to meet workload requirements, without rebooting the server. Because of the limitations inherent in Ethernet, early-generation IaaS providers forced users to add compute resources by adding additional servers and scaling horizontally. Secondgeneration providers based on the InfiniBand technology, like ProfitBricks, offer customers the ability to make their existing servers larger. "ProfitBricks offers great support: its engineering team has answered all our questions and solved our problems in real time,” says Khasanshyn. “We look forward to implanting future platform upgrades. They’ve been a great partner."
The Partner-Friendly Cloud Computing IaaS Provider
With channel-friendly reseller and referral programs, ProfitBricks makes integrating a powerful big data project into the public cloud simple and profitable; while meeting project timelines and exceeding performance requirements and cost targets.