By: David Wilkie – Technical Solutions Architect | March 18, 2016
Software Defined Data Centers are on the minds of technology folks everywhere: ‘Where does it fit in?’ ‘How does it fit in?’ ‘What are my options?’ PDS has featured Software Defined Data Center presentations and blog posts in the past, however what I want to do today is talk about another Software Defined option called Hyper Convergence.
Hyper Converged systems consist of compute, storage, and networking wrapped up into a few U’s of Rack space. These systems are defined by how easy they are to implement, and how easy they are to scale out resources. Hyper Converged systems also tout built-in features to make administration easy and reliable. Data Protection, High Availability, and wizard-driven deployments to name a few.
While we heard about recent start-up ventures that entered into the Hyper Converged market, we sometimes forget that the established entities and technologies of companies, like HPE, already had the foundation in hardware and software to really up the ante of Hyper Convergence. HPE has done this with the HC250 platform developed for virtualized workloads.
So what does one of these systems look like? Let me introduce you to Hewlett Packard Enterprise’s HC250 Hyper Converged system. A picture is worth a thousand words, so here it is:
I know what you’re thinking… This is one great-looking hyper converged system from HPE, but can it stand up to the competition?
If you guessed yes, you are right. The HC250 combines existing knowledge and new innovations to leap frog the competition in hyper convergences. HC250 is built on the Gen9 3 or 4 node ProLiant Apollo server platform and mature StoreVirtual software defined storage. This means you are already familiar with the server and storage platform from HPE because, chances are, you have used one or both solutions in the past. This 2U form factor system can scale to 8 nodes in a cluster with 4 servers in each node. This is a turnkey system designed from the ground up, based on a solid foundation, to provide an SDDC appliance that will fit several use cases (we will get to those in a minute).
First, what is in the box?
- 3-4 HPE Apollo Servers per node.
- VMware vSphere or Hyper-V (requires licensing)
- HPE StoreVirtual storage software
- HPE OneView for VMware or Microsoft System Center
- HPE OneView InstantOn for 15 minute deployment
- Factory Installation of all software weather you perform an HPE config to order (CTO) or not
Second, what are the use cases?
- Primary virtualization platform for medium size business
- Dedicated resource pool for specific application in enterprise (think VDI, SQL farm, Oracle farm)
- Robo/Branch application (think directory services, File/print Services, Point of sale)
- Test and development
Third, what are the technical details?
I am glad you asked. This solution utilizes wizard-driven software for a simple guided system setup. You provide an IP, credentials, and licensing and you are deploying virtual machines in 15 minutes or less! You manage the HC250 through vCenter or Microsoft System center depending on your virtualization platform. Here at PDS, we love both. Keep in mind all the software: storage, hypervisor, and management software are all ready to go so you do not have to deploy that.
The technical bits:
- 2U form factor with up to 4 nodes per system
- 1-8 systems per resource pool
- 128, 256, or 512 GB of memory per node
- Intel Xeon E5-2600v3 family of processors
- 6 drives per node for a total of 25 SAS and/or SSD drives (hybrid or all flash)
- 11.6TB max usable space (drive dependent) in clustered storage
- Network RAID 0,5,6,10+1, and 10+2 (Different RAID level require different configurations)
- HA and DR capabilities
- Tiering SAS and SSD for efficient storage
- ILO remote management
- 1 or 10Gbe networking options
The HC250 family employs StoreVirtual SDS (software defined storage), based on the trusted LeftHand OS, to drive thin provisioning and network RAID for data availability. HPE goes beyond data availability and introduces data protection. You can use application consistent thin snapshots that quiesce virtual machines. These snapshots are point-in-time copies of data on a per-volume basis. Administrators access snapshots to recover individual files from the volume, or they can roll back the entire volume. StoreVirtual’s Remote Copy feature allows replication to offsite servers with VSA data protection shield to ensure business continuity. You can even point your snapshots to a StoreOnce purpose-built backup appliance to move your data protection to a secondary storage device.
The HC250 is also vSphere Metro Cluster certified! This means you have stretch cluster capabilities to support business continuity by keeping applications online during appliance, Rack-level, or site-wide outages.
You can see HPE is not a newcomer to Hyper Converged Systems, in fact, they laid the ground work years ago to build the most complete Hyper Converged System. By turning to HPE and PDS for your Hyper Converged implementation and service needs, you are utilizing experts and tried-and-true technologies to deliver on your business needs.
Please feel free to reach out to your Account Director at PDS or email me at email@example.com to get more information.
David Wilkie is a technical solutions architect for PDS that concentrates on data center design. David is responsible for architecting Virtualization, Server, Storage, and BURA solutions. David has worked for PDS for 5 years and can be reached at firstname.lastname@example.org.