#OpenStack is a bunch of software’s pieces, right?
Besides of #OpenStack brings so much resources to cover most of your needs to get a really enhanced cloud experience. You need to add more pieces and coding to fully fit your performance needs.
Choose the right Hyper-Visor
Let’s start choosing the Hyper-Visor. You can get KVM, Xen, Hyper-V or VMWare. Any of these options are OK. However, if you have plans to run SAP or Oracle on top, then your list is smaller – to not say there is just one option left –
Mostly web portal are moving into #OpenStack solution and they can be developed in opensource platform, this brings you more options. Therefore, depending on what applications are you thinking to run on top your private, you have to define what Hyper-Visor fits better.
Also, you need to consider the level of integration that this chosen Hyper-Visor has with #OpenStack, and the experience of your DevOps.
#OpenStack doesn’t bring a easy 100% integration based on our experience with any of them.
For example, Is “live migration” from Hyper-Visor fully controlled by #OpenStack?
Choose your SDN strategy
There are options provided by different vendors. However, there is a risk about getting a vendor lock-in in your private. Of course you can choose #OpenVSwitch and avoid this issue.
Anyway. VMWare NSX helps you to work with any hardware network vendor, but you need to pay for it and of course. NSX is a terrific solution and you don’t regret to go with this one. On the other Hand. #OpenVSwitch project – where VMWare is one of the most important code contributors – is free of charge, but it requires talent to make it work out, it doesn’t bring a seamless integration with OpenStack. More details about #OpenVSwitch at “there’s real magic behind openstack neutron“.
A big issue for SDN is that you really need to do something to control network traffic from every instance. There are strong SDN solutions that provides management tools to rid of “Noisy Neighbors”… or maybe you have the talent to implement solutions like Intel SAA.
Choose your SDS solution
Well, this is not an easy journey. I strongly suggest read “what do i have to choose to efficiently store, protect and manage my private cloud’s data?”
We have chosen #Ceph because we were looking for something that could bring a high performance based on SSD drives with lower prices, and with fully integration to #OpenStack Cinder. We was challenged to make it work, but it worths every minute of our time. We’ve got a really nice performance and not operational issues… until now.
Of course, the market offers more solutions like Solidfire, EMC, NetApp… But, depends on how much talent your DevOps could pour into it, the level of help you requires from these big boys.
Our solution #KoolFit is based 100% on opensource and as you could read, you need talent to keep it up and running… and changing… Our mission with #KoolFit is to bring an affordable #OpenStack private solution with enough redundancy, automation and performance to host your most dynamic business applications.
DevOps is the most important resource. They will add or remove pieces of software depending on how the service roadmap and customer needs could change over the time. The today’s snapshot is shown in the following list:
- Ceph – This is the storage system where the compute instances run on
- Zabbix – We use it as the monitoring system for Cloud components and compute instances
- MariaDB – Database where our cloud related data and configurations are stored
- Nagios – We use it as the monitoring system for Cloud components and compute instances
- Openvswitch – Software-based Switch for compute instances communication
- Dnsmasq – DHCP leasing component for compute instances
- HaProxy – Load Balancer for compute instances
- Qpid – Data Messaging component for communication between cloud components
- Libvirt – Virtualization API to interact with different hypervisors
- KVM – Hypervisor where compute instances run on
- Apache HTTP Server – Web server for web pages and dashboards
- Nginx – Web server for web pages and dashboards
- MongoDB – Database for metering and usage data for cloud environments
And we are sure this list will grow much longer as we add more services and features to #KoolFit.
Stay close and see how we will change the market from this part of the world – sounds confident, right? well, we are –
See you around!