Hardware dedicated servers with a primary function to run software based web servers like Apache, Microsoft IIS or NGINX, which all serve content to the World Wide Web. Depending on the type of application or website, you are running and the amount of its users you will need a more or less powerful hardware. PhoenixNAP dedicated servers give you full control and customization of hardware including different storage options, RAM speed and capacity as well as latest Intel CPU’s with a wide range of clock speeds and core counts.
Servers dedicated to running database server software like Microsoft SQL , MySQL, PostgressSQL, MongoDB etc. Depending on the sizes of your databases and the amounts of simultaneous users data base applications can be extremely demanding and require many hardware resources. Dedicated servers meant for running heavy databases should be equipped with dual Intel Xeon CPU’s, healthy amount of RAM and local storage with high IOPS like SSD and/or NVMe drives. For fast caching purposes, a combination of NVMe and SSD is a most cost effective solution.
Hardware meant to support VPN services as its primary goal is usually less demanding in terms of capacity while more important role here plays the location and thus not only proximity to end users, but also very often the idea of being in a certain geo region for connectivity purposes. Along these demands, the bandwidth speeds and costs play a vital role in such deployments. PhoenixNAP large geographic spread offers you access to global strategic physical hardware locations and includes large bandwidth packages to support your deployment goals.
These servers primary function is to provide a large capacity with varying amounts of IOPS depending on the type of storage you are building. Our dedicated servers meant for storage purposes vary from large capacity 16-bay systems able to house up to 10TB HDD’s and/or 1.9TB SSD’s to 8-bay system supporting NVMe storage and various sizes of SSD/HDD drives. These systems are ideal for building very fast storage nodes, tiered storage systems as well as cost effective large capacity storage options for backup repositories. Reliability of our enterprise grade hardware can be easily extended with RAID 5, 6 and 10 options. On top of that we offer you access to our own fully S3 API compatible Object Storage solution to give you even more storage options, limit costs and take advantage of turnkey solutions.
Performance and latency are of great importance for the users of these systems and that is why our dedicated game servers come with wide range of enterprise grade hardware including plenty of CPU choices depending on your software needs supported by fast SSD local storage and a global distribution and availability. On top of that, we throw in large bandwidth packages and all that comes with a complimentary DDoS protected network. High clock speed CPU’s with 4.10 GHz (4.50 GHz Turbo), high IOPS local storage and our reliable network will keep your games running and your clients happy.
Virtual environments usually require multiple and multicore CPUs, large amounts of RAM and decent IOPS from storage. Using dedicated servers for virtualization purposes would lead you on a path of looking at systems with dual CPU’s with 12 or 16 cores each, a healthy amount of RAM to support your Virtual Machines performance needs. Local storage performance requirements can be achieved with the use of NVMe and SSD drives. Ultimately robust virtualization platform builds will require a multitude of servers and shared storage with availability and performance in mind. High Availability features built in hypervisors and RAID 6 and 10 storage builds will protect your clusters from hardware failures, even though we include 4 hours break fix SLA’s you might not be in a position to have any downtime. Our 10G network options, DDoS protection and 100% network uptime SLA’s will take care of your connectivity requirements.
Staying in control of your data is an extremely important factor of any deployment whether you are trying to limit and control or even completely block access to certain systems within your deployment from the prying eyes of unauthorized users your first line of defense, monitoring and control of your network are dedicated firewalls. Our firewalls options will allow you to fully control and manage your network traffic based on your own security rules. On top of firewalls, you can support your website and domain with SSL certificates, which will not only help with traffic encryption, but also help you prevent penalties by Google search ranking as an untrusted site.
For all the transcoding tasks whether you use Intel Quick Sync with our E3-1275v6 or E3-1585Lv5 (with Iris Pro p580) systems or general purpose GPU like NVIDIA Tesla P40, or even high end Intel Xeon Scalable Platinum series processors, at the end of the day you are trying to achieve certain quality of streams using average numbers of concurrent streams per machine. Ultimately you want to achieve the most economic scenario in terms of cost per stream and be able to support x-amount of concurrent users with certain quality. There is a large number of moving factors here and having access to customizable and on demand hardware deployments is key. Our dedicates servers options will support all your streaming needs for video compression standards like H.264/AVC, H.265/HEVC or VP9 across multiple strategic edge locations and supported by a high quality network with excellent bandwidth consumption packages.
Virtual Desktop Infrastructure (VDI)
Remote access for graphic accelerated desktops for workers anywhere and at any time is the core idea behind these types of solutions. Our dedicated servers equipped with high performance Intel Xeon Scalable CPU’s with support of NVIDIA Tesla P40 GPU’s, fast SSD/NVMe storage and large amounts of RAM are able to support any VDI deployments from Windows 10 desktops to complex engineering and architects projects on a truly global scale. Ability to virtualize application and use large datasets for rendering and simulations with global access and added security through access and data residency control paired with our high performance network and free DDoS are there to improve your business agility.
High Performance Computing (HPC)
The use of parallel processing algorithms to solve more complex computational problems in short amount of time is one of the main goals of HPC. Health and life sciences, chemical & mechanical engineering, geosciences, weather and financial modeling are some uses of HPC today. These computational models require access to a large multitude of cores which today can be achieved not just with larger scale of high end CPU’s, but high performance GPU’s like NVIDIA Tesla V100 are excellent for these types of jobs. Parallel computing with high end GPU’s is also massively adopted in Big Data deployments. Access to thousands of cores from each GPU with high capacity network and high IOPS storage allows for ideal infrastructure builds for HPC and Big Data applications deployments alike.
Delivery of AI services became increasingly important for today’s enterprises and deep learning deployments became strategically important for many businesses. The ability to leverage AI for predictions, classifications and analytics make a huge difference for any business with access to large and complex and ever growing data sources. Deep learning training and inference at the core of delivering such AI services. The ability to achieve short training times for Deep Learning projects is key, especially since deep and complex neural networks with massive data sets can take months to train if you don’t have access to the right hardware resources. PhoenixNAP bare metal servers with multiple NVIDIA Tesla V100 GPU’s put thousands of CUDA cores at your data scientist’s fingertips. This innovative infrastructure with use of popular frameworks like TensorFlow, CNTK, Caffe, Keras, PyTorch, Theano and others will enable achieving much shorter training speeds vs. any deployments with general purpose CPU’s or even older architecture and generations GPU’s. On top of that our Tesla V100 and P40 GPU’s offer excellent support for inference deployments so you can really put to work all the training models and achieve great performance and cost savings for your AI production workloads.