Local vs Global, part 3: How ready is your business for the hyperscale cloud?

By Arthur Shih

We’ve discussed the key differentiators, and key benefits of local and hyperscale clouds, but let’s now take that final step to see how ready your business is to move to a hyperscale cloud.

Step 3: So how ready is your business for hyperscale clouds?

The answer to where to host your applications depends on the answer to a couple of key questions (choose A or B).

1. What does my application do?

a. Line of Business – my business uses it to generate income
b. Test/Development – does not directly generate income for the business

2. What happens to my business if the application is unavailable?

a. The business will not run
b. We can survive on manual processes for a little while

3. Where are my users located?

a. Primarily in New Zealand
b. Primarily Overseas

4. How does the application link in with my other systems?

a. Primarily via local network access
b. Primarily via APIs and Web Calls

5. What consistency does the performance of my application require?

a. Must have consistent performance at all times
b. Can accept variations in performance at certain times

6. Where does my business data have to reside?

a. Must be in New Zealand
b. Can be held anywhere

7. How important is knowing how much hosting the application will cost me each month?

a. Must know how much each of my applications will cost me each month
b. Can deal with variable costs each month

8. Does the licensing terms of my application allow me to host in hyperscale clouds?

a. My applications are only licensed to run on my own kit
b. Application licensing allows for hyperscale cloud hosting

9. How well does my application respond to latency?

a. Application latency is not an issue
b. Application latency is an issue

If you answered ‘a’ to the questions above, 4 or more times, then your application would be more suited to a local cloud deployment. If you answered ‘a’ only three or fewer times, then your application is possibly suitable for a public cloud deployment.

Arthur Shih is Datacom’s cloud solutions manager.

Local vs Global, part 2: What’s the difference between a local and hyperscale cloud?

By Arthur Shih

We are still on the journey to discover if your business is ready to move to a hyperscale cloud. While we discussed the key differentiating factors in our previous blog, let’s take a look at the key benefits.

Step 2: What are the benefits and differences of local and hyperscale clouds?

Design for Failure

The main difference between the local clouds and the hyperscale clouds that users will experience is the concept of “design for failure”. The local clouds utilise enterprise software and hardware, creating “bullet proof” environments that are protected from hardware failures, meaning you can just move your environments and always be protected from hardware failures.

Hyperscale clouds use more commodity hardware. Individual virtual machines and workloads do not have guaranteed uptimes at all, and when something fails, the whole virtual machine will fail. SLAs for hyperscale clouds typically only guarantee availability if you have two instances of everything – otherwise they offer no guarantees. This is where the concept of design for failure comes into play. You will have to ensure you design around the availability limitations of the hyperscale clouds.

Having said that, let’s take a look at the benefits of each of the cloud platforms.

Local Clouds

Migrate workloads “as is” –  Due to the nature of local clouds being built on enterprise infrastructure that protects itself from failure, workloads and applications can be migrated into them without any re-architecture or design. Simply pick up your current environment and deploy it on the local clouds and you are away.

Customised managed services –  Local clouds are by nature able to offer a more granular and tailored SLA. Most hyperscale clouds offer SLAs based on the virtual infrastructure of a customer environment, but never the operating systems or applications within the infrastructure. Local clouds are able to offer guaranteed agreements for the restoration of full end to end services for customers.

Fixed price – Local clouds generally offer guaranteed infrastructure at guaranteed prices. Once purchased, the infrastructure is there for you to use without any extra costs – there are no extra bandwidth utilisation or storage input/output charges to access something you have already purchased. Hyperscale clouds, while offering unlimited flexibility, also offer the risk that costs may blow out in scenarios while scaling out automatically or in an uncontrolled fashion.

Local contacts and management – The team providing your managed services on top, and your cloud infrastructure are typically part of the same company. This means that when there is an issue that affects your critical business application, local cloud providers are wholly responsible for delivering your application through to your customers. There is no need to contact large out of country corporations, and no confusion over accountability.

More flexible licensing deployments – One of the show stoppers for organisations going to cloud computing is around the licensing of their applications. Many times organisations are only licensed to run their applications on their own equipment and not on shared equipment (e.g. customers without Software Assurance). Local clouds are more flexible to “pinning” customer virtual infrastructure to individual pieces of hardware, catering to strict software licensing rules while offering customers the benefit of high availability and scalability that clouds can provide.

Certainty over data ownership and recovery – Local clouds are very flexible when it comes to retaining the ownership of data that you put into their clouds, allowing you greater freedom when it comes to retrieving your data in the event of a contract termination. Hyperscale clouds often reserve the ability to view, modify, or move your data without your knowledge. Some hyperscale vendors have clauses within their contracts giving them the right to delete your data (most of the time without prior warning) if they deem it to be “suspicious.”

Know where your data is at any time – Local clouds are typically more transparent, and by their very nature give you the ability to understand the exact location of your information, even allowing you to visit the data centres where they are located. Global cloud providers are less transparent about the location of your data, typically only advising you of the regions they are in, with terms that allow them to move the data as required.

Avoid exchange rate risk – Local clouds transact in local dollars, meaning that your month to month spend is predictable and understood. When purchasing services from global clouds, transactions normally occur in foreign currencies, which is subject to unpredictable fluctuations, making your bill unpredictable. A five cent move in USD – NZD or AUD conversion rates will mean an extra $10,000 cost per annum on a $10,000 per month bill.

Hyperscale clouds

Faster to deploy – Hyperscale cloud deployments are all internet based, with no solution design required. This allows customers to deploy infrastructure within minutes.

Lower cost – Hyperscale clouds provide very little flexibility in their deployment, and very broad SLAs, meaning they are able to offer what seems to be cheaper infrastructure costs.

Larger number of development tools – Hyperscale clouds were initially developed to cater towards start-ups and the development market, and as such have plenty of tools and APIs catering to automatic deployments and quick provisioning. Local clouds are more catered to the enterprise market, with more focus on the traditional design and build processes.

Unlimited and instant scalability – Hyperscale clouds have much larger deployments, and as a result have much more scale than what a local cloud provider may be able to offer. This allows global cloud vendors to offer much larger flexibility for scaling resources up and down. Local clouds, due to the realities of having smaller scale deployments, may have restrictions on the ability to scale up and down. This is also sometimes reflected in different service termination rights and obligations.

We’ve taken a close look at the differing benefits of local and hyperscale clouds. In our next blog we’ll ask the right questions to determine whether your business is ready to move to a hyperscale cloud.

Arthur Shih is Datacom’s cloud solutions manager.

Local vs Global, part 1: Is your business ready to move to a hyperscale cloud?

By Arthur Shih

In this blog series we take a 3 step look at understanding whether your business is ready to move to a hyperscale cloud.

We’ve all heard the mantra, cloud computing brings with it the promise of enabling organisations to consume IT infrastructure resources in an increasingly cheaper way, while providing unlimited capacity and scalability on a pay-as-you-use model.

However, this uncapped capacity and scalability depends on whether your business is ready to tap into a global cloud services model or alternatively opt for the use of a local cloud services model, both with differing benefits that may meet your business requirements.

With global cloud vendors such as Amazon and Microsoft being hosted out of regional centers such as Singapore and Sydney, latency and sovereignty issues have presented a barrier to adoption for New Zealand businesses. As a response to this, local cloud providers, such as Datacom have successfully offered their Datacom cloud services to the New Zealand market.

Step 1: Choosing a local or hyperscale cloud

As a general rule, the main differentiator offered by local clouds over hyperscale clouds, other than geographical location, data sovereignty and the ability to wrap custom managed services on top of the provided infrastructure are;

  • Local clouds due to their smaller scale and understanding of the local market, can offer much more granular and flexible arrangements to suit local businesses, but potentially with fewer features and at a higher cost than hyperscale clouds.
  • Hyperscale clouds on the other hand have access to more scale and therefore can often deploy more quickly and cheaply with more features, but less flexibility around guarantees and actual service management.

Added to this difference is the fact that local clouds are able to offer customers full end to end SLAs based on the outcome of the applications being hosted.

It is important to understand that the infrastructure purchased from global clouds is purely infrastructure only, with no guarantees around the specific application or service you are deploying on top of the infrastructure. They key is to understand how to design and deploy your applications in a way that takes into account the limitations and failure scenarios that may occur.

Hyperscale clouds on the other hand benefit from having larger scale and are able to offer scale up and scale down capabilities with lower price points than local clouds.

Larger scale also means they can offer more, though less flexible, functions and services at the infrastructure level, providing faster routes to standing up bare infrastructure.

Global vendors such as Amazon’s move to Australia with AWS, have resulted in customers starting to question whether or not their applications and systems can be hosted on Amazon and/or other global clouds. Datacom has been watching this space with interest, and to help customers understand their options, Datacom has:

  • Joined the Amazon Web Services Partner Network as a Consulting Partners. With over 30 trained technical staff in region, making us one of the most highly trained AWS partners in region.
  • Joined the Microsoft Cloud OS Rapid Deployment Programme for deployments on Microsoft Azure (Azure)

We have successfully helped multiple customers transform their businesses using both Amazon Web Services and Microsoft Azure, as well as using our own Datacom cloud services offering as well.

A summary of the key differences to consider when comparing local and hyperscale clouds:

We’ve looked at the key differentiators. In our next blog let’s take a look at the key benefits of local and global services models.

Arthur Shih is Datacom’s cloud solutions manager.