Azure Dynamic Scaling Update

In an earlier blog, I highlighted the lack of autoscaling features in Azure as compared to Amazon’s CloudWatch.  This left architects with the choice of a DIY approach using additional tooling (such as WASABi or MetricsHub) or using a 3rd party product – Paraleap’s AzureWatch. However, this summer Microsoft introduced an Autoscaling feature which is built-into Windows Azure directly – Windows Azure AutoScale (WAAS).

Microsoft Vice President Scott Guthrie recently announced in a blog post, “AutoScale enables you to configure Windows Azure to automatically scale your application dynamically on your behalf (without any manual intervention) so you can achieve the ideal performance and cost balance. Once configured, it will regularly adjust the number of instances running in response to the load in your application,” Guthrie wrote.

Guthrie also noted that WAAS supports two load metrics — CPU percentage and storage queue depth, for cloud services and WAVM’s only — but Microsoft will continue to add more services. Enterprises can set up WAAS in the Windows Azure Management Portal’s new Scaling page – see Figure 1 below.

scalesetting

Figure 1 – AutoScale Configuration Page

The method of operation for the Scale by CPU algorithm is shown in Figure 2 below.

Azure Autoscale by CPU

Figure 2 – AutoScale by CPU Algorithm

This feature is still in Preview mode rather than general availability but further enhancements are promised as Microsoft play catch up to market leader Amazon AWS.

Posted in Cloud Computing | Leave a comment

Project-Based Funding: An Obstacle to Private Cloud

The traditional physical data centre consists of a hodgepodge of equipment, tools and systems— all driven by project-based funding. This budgeting anachronism is an anathema to a virtualised data centre that utilises shared resource pools and can absolutely derail a planned transition to a private cloud model. Read the White Paper here.

Posted in Cloud Computing, Cloud Strategies | Tagged , , , | Leave a comment

Dynamic scaling in the Cloud

One of the key criteria for Cloud Computing, according to the NIST definition, is “Rapid elasticity”, i.e. “Capabilities can be elastically provisioned and released, in some cases automatically, to scale rapidly outward and inward commensurate with demand”. The reference architecture for this can be found in Cloud Computing Patterns under the heading of Elastic Component – see Figure 1 below.

elastic_component1

Figure 1 – Elastic Component Pattern

The solution is based on a component called a Scaling Manager which monitors the resource utilisation in real-time and, based on a set of metrics and rules, automatically provisions or de-provisions capacity from the resource pool.  A good reference implementation for this is Amazon’s CloudWatch shown in Figure 2.

CloudWatch-Overview

Figure 2 – Amazon’s CloudWatch

Unfortunately, Microsoft do not provide an equivalent in Azure – they just provide the API’s and toolkits for a DIY approach. However, this is where the Microsoft Partner Ecosystem comes into play. Paraleap Technologies is a Chicago-based software company, focused on providing tools and services for cloud computing technologies. AzureWatch is Paraleap’s flagship SaaS product, designed to add dynamic scalability and monitoring to applications running on Microsoft Windows Azure cloud platform. The architecture is shown in Figure 3.

azurewatch-architecture-diagram4_thumb

Figure 3 – AzureWatch

Further details can be found at http://www.paraleap.com/AzureWatch.

Posted in Cloud Architectures, Cloud Models, Cloud Patterns | Tagged , , , , | Leave a comment

Step by Step Guide to Office 365

This poster-style graphic provides a very good overview and summary of Office 365…

Microsoft Office 365 Guide
Infographic by: Cloud Hypermarket

- Posted using BlogPress from my iPad

Location:Union Drive,Birmingham,United Kingdom

Posted in Cloud Computing | Leave a comment

Journey to the Cloud – it’s not rocket science!

To me, as an Architect, the Cloud has become a “no-brainer” – it’s just business as usual. People often cite security as a barrier to entry but this perception assumes that “Cloud=Public”, however, there are other cloud models including Private Cloud, which can be on-premise (behind the corporate firewall) or hosted by a third-party. Some also say that existing applications may be incompatible if they are not cloud-aware and that may be true for Platform-as-a-Service (PaaS) implementations but there’s still Infrastructure-as-a-Service (IaaS). If an application runs on virtual servers like VMware (which we’ve been using extensively over the last five years or more) then why can’t they run in the Cloud? By the way, did you know that you can run your own Amazon AWS compatible Private Cloud on Eucalyptus Open Source software?

When looking at the options for cloud deployments (versus the out-dated on-premise options), Solutions Architects should look at the various architecture and design patterns such as those found on Cloudpatterns and Cloudcomputingpatterns.

For Architects who need some platform specific reference architectures then take a look at AmazonAWS and AzurePatterns. For their associated Visio stencils which can be used for diagramming these architectures and patterns then they can be downloaded from AWSIcons and AzureStencil respectively.

Link | Posted on by | Tagged , | Leave a comment

Is the Government’s CloudStore confusing the market?

As the UK Government iteratively develops its G-Cloud multi-source procurement frameworks, now into their third phase with “G-Cloud iii”, it seems that the messages are somewhat confused. Just because a supplier and service is available via the CloudStore doesn’t necessarily mean that it is accredited or certified to a particular security level. For example, I’ve seen public references to services like those from Memset as being IL2 accredited but what does that mean? Reports suggest that their “differentiated” IL2 service is available via the “public  infrastructure” – well, from their website descriptions, they have dedicated servers in their own data centre(s) so I assume this means access via the Internet although it’s not clear whether this requires some form of VPN or SSL for encryption purposes. I would have thought it should be accessed via the Public Services Network (PSN) which is, by definition, accredited to IL2.

In fact the UK Government’s ICT Strategy is based on the architectural trichotomy comprising the:

  • G-Cloud;
  • Public Services Network (PSN); and
  • CloudStore or Government Application Store (G-AS).

The overall architecture is illustrated in Figure 1 below.

G-Cloud Architecture

Figure 1 – High Level Architecture of G-Cloud

As such, the G-Cloud is a private cloud which is available only to the UK Public Sector community (which includes wider government, i.e. the Third Sector). Any SaaS applications that sit outside in the Public Cloud, such as Salesforce.com or Microsoft’s Office 365, should therefore be accessed via a secure gateway service if architectural integrity is to be maintained. This is where Integration and Management Services come into play!

CloudStore is an online procurement portal through which buyers can call-off services from the frameworks but these and the architecture are still evolving. Until we get more clarity on the target architecture and roadmap its going to be a bumpy ride.

Posted in G-Cloud | Tagged , , , , , , , | Leave a comment