The Cloud Has Burnt Me In the Past Can I Trust It Now – Part 2

In Part 1 we discussed the first 4 items affecting successful cloud trust. Today we will go through the remaining 4 and I’ll give you some recommendations. Let’s ensure they won’t bite you and your business when you to move to the cloud.

5) It is just too far away to trust

Trust DistanceDepending on where you are and what service you choose with cloud provider XYZ this has been the cause of significant headache for businesses I speak with. Often not considered until it’s very much too late.

Some providers have data centres dispersed across the country whilst others will only have a single presence. Often they may have a presence near your business but not every service they offer will exist in the country. The service you require may indeed run in another country even though there is a local data centre presence.

This brings us to the actual problem which is one of latency. Does the application you are running work as it should be hundreds of km away? Is there a requirement for it to be closer for the best user experience? If parts of the application are still housed locally and they communicate with one another often is the distance an issue?

Recommendation

If you are choosing an application that has more than one part then ensure that all parts go to the cloud provider. At the very least ensure that the cloud provider has a data centre within a few milliseconds of latency.

If it is a standalone web application then latency is not typically a concern. Some applications user base is dependent on high speed low latency access. In that case ensure that you use a cloud provider within the same region otherwise users will have an unpleasant experience.

6) Backups and DR weren’t thought of

When speaking with customers about what they’re using for backups or DR the answer is often “Oh I’m with cloud provider X”. This is because of the assumption that the cloud provider provides all this by default.

This is never the case and unless specified otherwise it will only ever exist as a single instance. Often it will be a highly available and robust instance but nonetheless reliant on a single data centre, storage device or switch.

In the event the data centre or region goes offline (and this happens all too frequently) customers are typically just left with downtime. They hear nothing until the provider comes back online or resolves the issue. This could be hours, days or even include data loss.

Recommendation

Have a talk within the business first about how highly available you need your cloud application to be. Discuss how much data loss you can afford before it starts to affect your business negatively.

With the information on how resilient you need your application to be and how much data loss you can afford, speak with your cloud provider. Ask what is included by way of backups and DR in the base product. Then make the necessary additions to include enough added functionality to give you the backup and restoration ability you require.

7) Not as resilient as your current environment

It really irks me that customers often equate the marketing budget of a cloud vendor to its capabilities and maturity of their offering. The two large hyperscalers have only being in Australia since 2012 and 2014. The maturity of their offering locally often does not compare to those that might have been operating for many years longer.

What this often will mean for businesses is that features they take for granted like high availability or load balancing of resources (HA and DRS) simply don’t exist. They trust that since they can lose a host in their own environment and the workload will power up on an alternate host that the same will be true in the cloud.

This is not true! Many cloud providers will leave your application down entirely until the host comes back online…

Recommendation

When choosing your cloud provider, find out what platform is it running on. Is it Hyper-V, VMWare, KVM, OpenStack? Once you know what the underlying hypervisor is, find out what high availability and resource levelling options your applications will benefit from.

Can you trust it handle the host going down? Can it be trusted to handle a noisy neighbour on the same host? Do you have to pay extra for this functionality that most assume is a basic feature that should be included?

Ask questions before you find out the hard way.

8) Loss of actual functionality

Enormous amounts of effort go into choosing the right hardware and software for your business. What you’ve chosen is going to be the best fit for your businesses requirements.

A customer I had a chat with recently found out the hard way. The firewall device they were using in the cloud was drastically outgunned. In terms of features and functionality it paled in comparison with their on premises firewall.

This meant that they simply were not able to do the firewall and load balancer rules that wanted. They needed these in order for their application to work effectively.

Recommendation

When selecting a cloud provider find out if they can give you a firewall or edge device that does exactly what you require. If your company needs to use BGP then make sure the cloud provider can give you an edge device capable of it.

Some cloud providers can’t provide you an adequate device. Find out if you can host a virtual appliance you trust to be used on the edge.

If your business requires a physical device to be collocated then is the cloud provider capable of doing this? Can they stitch it through to your cloud applications?

Some can, most cant so make a point of asking up front.

Conclusion

Yes, the cloud has burnt many of us in the past but most definitely we can trust the cloud now with our business applications/workloads. If you follow the recommendations in this article and ask the right questions that get to the root of what you expect from your cloud provider then you can choose the best fit for your business and not be yet another business “burnt by the cloud”.

Luke Brown

Author: Luke Brown

Luke Brown is an Advisory Systems Engineer at Dell EMC. He is passionate about all things tech and loves a good debate.

1 thought on “The Cloud Has Burnt Me In the Past Can I Trust It Now – Part 2”

  1. Not quite sure why the hypervisor actually matters. I would be more interested in my SLAs if I was a customer still.

    Understanding these SLAs and correctly architecting my apps is the critical part IMHO.

    Also keen to see you include info beyond IaaS like microservices, DBaaS etc in any review. The cloud is a chance for transformation, not just lift and shift (which also has its place)

    *Disclosure: work for MSFT

Leave a Reply

Your email address will not be published. Required fields are marked *