TheSaffaGeek

My ramblings about all things technical


Leave a comment

VCAP-CID Objective 1.4 – Determine Availability Requirements for a Conceptual Design

Knowledge

Identify availability options for management components.

  • Availability can achieved within the vCloud architecture in a number of different ways and via differing methods. I’m going to break them up into different categories and i’m not going to cover each one but if you understand the different methods I think when you are reading the vCAT or any other kind of design book you’ll be able to identify them with ease.
    • Redundancy: This is simply creating multiple instances of an important service to ensure that if one or more fail that the solution isn’t impacted by this. There are multiple examples of this but the most simple but one of the most important in my opinion is the creation and usage of multiple vCloud cells to ensure load balancing but more importantly redundancy in the event of a loss of a vCloud cell. You can also cover this further down the stack with Heartbeat in the vSphere layer(even though this has now been made end of life) ,multiple network cards from the physical networking and multiple redundant switches to multiple redundant storage processors.
    • Disaster Recovery/Failover: This is covered in a whole section in the vCAT which goes over methods of utilising products like SRM to configure disaster recovery of the management layer. For conceptual this is more about knowing what is and isn’t possible but also taking the availability requirements of the customer from a business impact analysis where it is deemed the amount of money a customer is willing to lose due to downtime and then equate this to a number of nines. The table below gives an example of the number of times compared to amount of downtime and with the larger the number of nines this will then mean more expensive solutions which you will need to advise your customer about (99.9 can be met by HA for example but 99.99 will require heartbeat and synchronise replication with QoS). For conceptual you don’t cover specific products but knowing that you will need a DR site with fast links between will cover this for example.

HA

Differentiate between management components and resource components.

  • This is simply determining what should be part of your management cluster and what should be part of your resource cluster. I think this is really straight forward as anything in your management cluster is used to provide services to you the vCloud administrator and the resource cluster/s are for your customers to provision to and is the pools of resources you configure as your provider virtual datacentres. The below image is a great example of a conceptual diagram of the management and resource clusters.

conceptual

Skills and Abilities

Explain compatibility of various vSphere high availability features with a vCloud design.

  • This is covered perfectly in appendix A of the vCAT Architecting a VMware vCloud pdf so I don’t see the need to explain it here and i think it is better if you go through that instead. The link to the online documentation centre is here 

Given customer requirements and constraints, determine appropriate customer Service Level Agreements (SLAs) for the conceptual design.

  • This is covered in more depth within objective 1.6 so we will cover this in that section.

Determine how given SLAs impact availability design decisions.

  • This is covered in more depth within objective 1.6 so we will cover this in that section.

Given customer requirements and constraints, determine how to achieve desired availability.

  • From the design workshops and requirements collecting you will have worked out what the customers requirements and constraints are and will then have to work with these to try meet them all. For this it is their availability requirements which will be as I mentioned above their permitted amount of downtime per year along with their RPO’s ,RTO’s , MTD’s and WRT’s. From this you will have to work with their constraints to design a solution that meets their requirements so for example if they have an RPO of 5 minutes for critical systems within the management cluster in the event of a site failure this cannot be achieved via SRM with vSphere replication. For the conceptual design my example isn’t applicable but knowing this kind of limitation will then mean you know conceptually what needs to be created (multiple sites with fast links that have near zero latency for multiple data service providers and storage that can achieve this)

Given customer requirements and VMware technologies, determine availability impact to the conceptual design.

  • I feel this is largely what i have mentioned above but now you are including VMware technologies limitations/capabilities into your thinking which I actually did above. You will need to know what is and isn’t possible with HA for example and how it’s can only provide a certain level of availability and is limited by the amount of restarts it can achieve at once whilst being possibly limited by priority groups.

If you feel I have covered something incorrectly please let me know as I’m learning like everyone else and I certainly don’t claim to be perfect (near it but not perfect Winking smile ). Also the vBrownbag covered the whole of objective 1 here.

Gregg


Leave a comment

VCAP-CID Objective 1.3 – Determine Capacity Requirements for a Conceptual Design

 

Skills and Abilities

Determine how storage and network topologies affect capacity requirements for a vCloud conceptual design.

  • This in my opinion can be taken in a few ways so I welcome any feedback on if you think I have looked at this the wrong way but the way I am looking at this is the way it is all connected to differing portions of the environment obviously impacts the speeds that can be achieved and thereby the capacity of virtual machines that can be run over a certain link for networking or even over a specific NIC/Switch/HBA/Cable. So to use the networking topology as the example:
    • Network: For networking there are a number of constraints that can affect the capacity requirements for a vCloud conceptual design. To give an example I will use one that I am seeing a lot recently which is a 10Gb NIC connection from each blade/rack server in your proposed vCloud environment. For this 10Gb link you need to carve it up (either via native hardware methods or via NIOC) for all the varying types of traffic that needs to go over the link for your vCloud environment. Now if your network topology is inside an existing datacentre then you may have to connect to an existing top of rack switch which may only have the capability to provide two 10Gb connections per switch and the price for 2 new 10Gb switches (to obviously provide resiliency) won’t fit in the budget. So for the conceptual design if you need 10Gb of network traffic leaving each host to supply network requirements of the virtual machines on the host then you will need to either:
      • Change the hosts to have a sufficient number of NICs to provide this or
      • Go down an infiniband route or
      • Explain to the customer due to the constraint of having to use existing switches it is not possible to provide the required network bandwidth for each host so they will need to buy more hosts so that the virtual machines on each host get their required bandwidth.
    • This way of thinking applies exactly the same for storage and if you are running converged networking then it can be almost exactly the same.

Describe VMware vCloud Director and VMware vSphere functionality and limitations related to capacity.

  • This in my opinion is all about vSphere and vCloud maximums which is always something you have to keep in mind when doing a conceptual design as for example the linked clone chain length limit is 30 and then after this a new shadow copy is created which then utilises more space on a new datastore and affects storage capacity. Actually knowing these functionality metrics and limitations is something I have been learning from going through the vCAT documentation. I did think about listing all of them but there are so many and what they could impact is so vast I think this is something where you need to know the limitations and functional capabilities of the two products and then think of it in the holistic manner of the whole design and how it impacts the conceptual design. Now remember the conceptual design is the “napkin” style design and so product names do not feature but you need to understand at a certain level what is and is not possible from the products.
  • As I mentioned in my previous point if you feel I am totally wrong then please do tell me in a friendly manner as I am certainly not perfect and am doing this to learn.

Given current and future customer capacity requirements, determine impact to the conceptual design.

  • During your design workshops you will work out and record what the customer’s current and future capacity requirements are and then will need to plan for that 20% year on year growth they require to give an example. So if their current requirements can be met with eight hosts to be very simplistic then you will need to ensure you have sufficient capacity not just in compute but also storage, networking, cooling, power and switching.

Given a customer datacenter topology, determine impact to the conceptual design.

  • For this I think I covered it in the first section but you now need to look at the whole topology with storage, networking, power, rack space, distances between components, distances between datacentres, cooling and weight limitations to name a few off my head that may impact your conceptual design. So say for cooling you can only put in a certain amount of hardware into each rack which then impacts your conceptual design of how many blades can fit into the datacentre/server room.

Given cloud capacity needs, constraints, and future growth potential, create an appropriate high-level topology.

  • This is the point where you have done your design workshop and are now looking to do a high-level design of the environment that meets all the customers’ needs and shows to them you understand what they require and have planned for the future. The below diagram is a very basic version of what you would provide based on networking to show you understand their needs :

    image