Friday, April 22, 2011

What we learn from Amazon Web Services cloud failure

 Learning from AWS  incident (main culprit seems to be the S3 design flaw) - we need to develop processes to easily have clouds on different providers and mitigate the risk that one of them fails.

Strategic Blue  "solution transforms multi-cloud infrastructure deployment from a potentially complicated management issue to a simple billing relationship with one party, Strategic Blue."

Applications like RightScale and Besol Tapp will move applications among different cloud providers. They providethe actual software the brokerage dealers in cloud resources need.

What should be strategy to select multiple providers? IMO, the the application requiring  MINIMUM effort to port on two or more cloud providers, are #1 choice for multi-provider hosting. The easiest would be to use Tapp or Rightscale.

We need a magic list of cloud providers customers may choose for multiple hosting for defense against cloud failure. I am paraphrasing Gartner Magic Quadrant for cloud providers

Monday, April 04, 2011

What's the difference between enterprise cloud and private cloud?

"Organizations going down the private cloud path have some tough decisions to make.  (Why Private Clouds Are a Catch-22 for Buyers, GigaOM)

This title question appeared on Quora . The first Catch-22 is to define what a private cloud is Here are some opinions, all valid.
"A private cloud is behind a firewall - the likelihood is therefore that it is being operated by an organization's own data center".  

"The term 'private cloud' as it's used now seems to miss the entire point of cloud computing."

" Enterprise Cloud is a euphemism that a small number of marketers and vendors will use to describe on premise blend of IaaS & SaaS. It is simply a private cloud plan and simple
"The private cloud means user own the resource privately and use the resource privately"

 "Private Cloud ... is a marketing term being applied to private IT resources that are organized, orchestrated, managed, deployed, and used in a service oriented model"

 The main reason any organization wants a cloud is to make more money than it does now. A bunch of computers, no matter how well orchestrated are not a cloud.  Here is my take on private clouds
"There is no difference between a private cloud and a customer using a public cloud. If you use your own physical servers, the creation of a cloud with elasticity & billing of resources is very complex, The private cloud software really relegates the CIO to the same headaches of running a data center today.

One can build exactly the same private cloud using software servers. IaaS providers like CloudSugma - give customers root access, and the option to install whatever software they need to mount the application and services, the same as for physical servers. You select your own security and firewall, or select it from the 3rd parties on CloudSigma. The software offered as PaaS is from 3rd parties only and it is optional. CloudSigma's customers select which 3rd party PaaS they like from CloudSigma site (jclouds, enStratuss, AppFrist and so on), or they can select none.

AWS slogan is "we do the heavy lifting for you'. This is true, the intention is to remove all complexity and so one can have it's own private cloud on Amazon public cloud. CloudSigma goes a step further, and gives customers the ability to create a cloud in a day, using the same experience curve familiar from having a bare physical server; load the OS, load the middle ware, use CloudSigma billing and ask your customers to use your services with a credit card.

Servers as software are better: cheaper, faster, more fault-tolerant, and more secure than traditional servers.  The shift from private to public cloud is related to hardware replacement cycle. So what we call today private clouds, will become simply a readily available service from public cloud. The owners of private clouds will have the same exclusive privileges they have on their own premises"
So why not build the private clouds using the public clouds resources?

A "pure IaaS" provider ?

Revolutionary,  yet evolutionary these thoughts will change the thinking on how to offer IaaS services to  customers respecting their dignity and wishes, while making them richer.
The largest Infrastructure-as-a-Service (IaaS) public clouds are increasingly morphing into more like Platform-as-a-Service (PaaS) operators.

The customers using cloud services, particularly at the infrastructure level, are also interested in other tools that may add convenience. These are usually service driven; things such as email services, content distribution networks, DNS services etc. etc. So the debate here really is who is best positioned to provide those services and how should they be offered/delivered to end customers?
Can Amazon Web Services (AWS) maintain a position where it offers the best email service, the best DNS service, the best database service …? The answer so far is yes

In all categories I can point to products and services offered today that in my opinion are better than the AWS alternative. Of course customers have different needs so the best for you might not be the best for another user. Don't customers deserve easy access to the best services for them? Surely the challenge for the IaaS cloud vendor is therefore to make the widest choice of services available easily to their customers, not to offer an anointed service or in-house offer.

CloudSigma describes itself as the pure IaaS cloud. This is a marketing metaphor, as 3rd party PaaS can create havoc, when not designed to work together.

One impact is to have a lock-in effect on the customer base (using proprietary  PaaS) to the particular cloud which is something customers might want to think about carefully. There are some even more worrying problems associated with the move to PaaS as the IaaS provider increasingly widens their role and comes into conflict with users of their own cloud
The primary motivations for IaaS cloud vendors to offer their own in-house PaaS services are that they feel they have a captive audience. If for whatever reason the customer wishes to move to another vendor, this is going to be a wholesale move of your entire operation. Whilst looking convenient up front, in fact creates a very entrenched situation for the customer which is difficult to escape from.

Vendor lock-in is not something desirable for customers (it is for providers) and should be avoided if possible when choosing a provider. Multiple providers is the solution.
Technology will keep re-inventing itself and evolving through creative destruction or if we'll see convergence and some abstraction from computing processes.
 In the US Netflix, the leading movie streaming service is estimated to account for some 20% of peak internet traffic! The company chose to migrate to the cloud ( most of their IT), and now largely uses Amazon's EC2 service. As a result, EC2 benefits from the bulk purchasing of data traffic that Netflix necessarily purchases from them in the course of its operations in their cloud.

Netflix are known to be happy with their cloud experience and Amazon clearly aren't complaining so what's the problem? Well, Amazon just launched a video streaming service which is a direct competitor to Netflix and their strongest competition currently. Netflix are hosting much of their infrastructure with their biggest competitor. Not only this but Amazon will likely be benefiting from lower data transfer costs over their whole operation as a result of the additional traffic that Netflix puts through their cloud! It is a neat trick but how do Netflix shareholders feel about it? :-) (Note December 2011; this paragraph was prophetic :-) )

Blog Archive

About Me

My photo

AI and ML for Conversational Economy