Buck Woody : Azure Use Cases, Computing

Creating a Distributed Computing System Using a Windows Azure Queue

BuckWoody — Tue, 11 Oct 2011 13:12:42 GMT

The Windows Azure Queue component, like all Windows Azure components (Roles, Storage, App Fabric, SQL Azure) can be used by itself or with other Windows Azure components. That’s why I refer to Windows Azure as “Distributed Computing” rather than “cloud”.

Having a distributed off premise queue has a lot of use-cases. An interesting use-case is a company that wanted to harness the power of all of the PC’s and laptops in the company when they were not being used throughout the day. A developer wrote a screen-saver program that connected to an Azure Queue, pulling work off of the queue and placing an entry when it was done. In essence he had a partially connected distributed work relay system, and since he used a Windows Azure Queue, the system worked from anywhere in the world.

He uses an on-site central server (which was actually only a workstation-level system) that holds the computations in a scatter/gather paradigm. The computations are broken into less-than-8K chunks, so that it fits within a message. The server connects to a Windows Azure Queue, and places the message marked for computation. It also scrubs the Queue for completed work, and as part of the process puts that kind of message into a mapping function (queues are not guaranteed a message order).

The workstations that are not being used (even those systems at remote workers and travelers) connect to the same Windows Azure Queue when the system is not being used for a period of time, when the screen saver kicks in. It then takes one message from the queue, computes the information, and then sets a new message for the server to pick up with the answer. The workstation then deletes the message.

The Server picks up the completed work, processes it and then deletes that queue message. He also added logic to process messages for computation on the server as well, when the server function of adding work is not required.

There are a few caveats here. This works because of the mapping function on the head server. Order is not guaranteed, so he includes a number for the function step as part of the message body, which cuts the size a bit. Also, he’s careful to watch the encoding, since Azure will hand binary back in Base64 format.

He’s found that there are enough systems to ensure that the messages are cleared every few days – important, since the Windows Azure Queue ages out after seven days. Also, he’s careful to use the CloudQueue.PeekMessage function when he wants to monitor the system – that function ensures that the message status doesn’t reset as “read” when he accesses it.

This is a great example of using the “cloud” as what it is intended to be – a distributed architecture you can use as needed to solve a business problem. It’s not an “all or nothing” proposition, but instead it is simply another set of components to use where you need them.

Rip and Replace or Extend and Embrace?

BuckWoody — Tue, 13 Sep 2011 11:20:05 GMT

As most of you know, I don’t like the term “cloud” very
much. It isn’t defined, which means it can be anything. I prefer “distributed
computing”, which is more technically accurate and describes what you’re doing
in more concrete terms.

So when you think about Windows and SQL Azure, you don’t
have to think about an entire product – you can use parts of the system
together or independently to accomplish what you need to do. You can use the
computing functions, storage, and more and more I see folks leverage the
Service Bus to enable current applications to expose things to the web.

And that brings up the point of this post. Once you decide
that a distributed architecture works to solve a problem, you’re faced with a
decision: should you completely re-write your architecture to take advantage of
the current systems or should you just fold in new code that makes the data or
function available to the web?

Of course, the answer is always “it depends” on the situation
– and it does. But unless you’re fixing a problem with current code, I usually
advocate a migration approach. That means at the very least retaining the
business logic (again, unless it’s not currently working) and as much of the
code as you can. In fact, if you follow this paradigm, you’re on your way to
making a Service Bus out of the functions you currently have. You can expose
the results of a system rather than opening the system up. Let’s take an
example.

Assume for a moment that you have an order-taking system
on-premise. That system performs many functions, one of which might creating a
Purchase Order. Your system might be enclosed, meaning that it has an
application that talks to a middle-tier, and then from there to a database
system. A query is generated from a screen, and passed along to eventually
compute, store and return a Purchase Order Number, along with other
information. Imagine now that you wire up the code not only to return the PO
number to the client, but to make that number available on an endpoint –
actually really not that hard to do.

Now you can make that PO number available to the web using
Azure. You could restrict who can make that call to the system, or open it up
to a broader audience. Or instead of the PO Number, you could make a product
list available. And you can go further than that – EBay, for instance, uses the
OData protocol (which is very cool in and of itself) which you can query from
the web. You could compare your company’s product catalog to what is on EBay,
and list the items you have there if there are no competitors in that space.
And on and on it goes.

So the point is this – where you can, retain what works.
Fold in systems like Azure where they make sense. Extend and Embrace.

Windows Azure Use Case: High-Performance Computing (HPC)

BuckWoody — Mon, 28 Feb 2011 15:52:31 GMT

This is one in a series of posts on when and where to use a distributed architecture design in your organization's computing needs. You can find the main post here: http://blogs.msdn.com/b/buckwoody/archive/2011/01/18/windows-azure-and-sql-azure-use-cases.aspx

Description:

High-Performance Computing (also called Technical Computing) at its most simplistic is a layout of computer workloads where a “head node” accepts work requests, and parses them out to “worker nodes'”. This is useful in cases such as scientific simulations, drug research, MatLab work and where other large compute loads are required. It’s not the immediate-result type computing many are used to; instead, a “job” or group of work requests is sent to a cluster of computers and the worker nodes work on individual parts of the calculations and return the work to the scheduler or head node for the requestor in a batch-request fashion.

This is typical to the way that many mainframe computing use-cases work. You can use commodity-based computers to create an HPC Cluster, such as the Linux application called Beowulf, and Microsoft has a server product for HPC using standard computers, called the Windows Compute Cluster that you can read more about here.

The issue with HPC (from any vendor) that some organization have is the amount of compute nodes they need. Having too many results in excess infrastructure, including computers, buildings, storage, heat and so on. Having too few means that the work is slower, and takes longer to return a result to the calling application. Unless there is a consistent level of work requested, predicting the number of nodes is problematic.

Implementation:

Recently, Microsoft announced an internal partnership between the HPC group (Now called the Technical Computing Group) and Windows Azure. You now have two options for implementing an HPC environment using Windows. You can extend the current infrastructure you have for HPC by adding in Compute Nodes in Windows Azure, using a “Broker Node”. You can then purchase time for adding machines, and then stop paying for them when the work is completed. This is a common pattern in groups that have a constant need for HPC, but need to “burst” that load count under certain conditions.

The second option is to install only a Head Node and a Broker Node onsite, and host all Compute Nodes in Windows Azure. This is often the pattern for organizations that need HPC on a scheduled and periodic basis, such as financial analysis or actuarial table calculations.

References:

Blog entry on Hybrid HPC with Windows Azure: http://blogs.msdn.com/b/ignitionshowcase/archive/2010/12/13/high-performance-computing-on-premise-and-in-the-windows-azure-cloud.aspx

Links for further research on HPC, includes Windows Azure information: http://blogs.msdn.com/b/ncdevguy/archive/2011/02/16/handy-links-for-hpc-and-azure.aspx

Windows Azure Use Case: Hybrid Applications

BuckWoody — Tue, 22 Feb 2011 19:44:09 GMT

Description:

Organizations see the need for computing infrastructures that they can “rent” or pay for only when they need them. They also understand the benefits of distributed computing, but do not want to create this infrastructure themselves. However, they may have considerations that prevent them from moving all of their current IT investment to a distributed environment:

Private data (do not want to send or store sensitive data off-site)
High dollar investment in current infrastructure
Applications currently running well, but may need additional periodic capacity
Current applications not designed in a stateless fashion

In these situations, a “hybrid” approach works best. In fact, with Windows Azure, a hybrid approach is an optimal way to implement distributed computing even when the stipulations above do not apply. Keeping a majority of the computing function in an organization local while exploring and expanding that footprint into Windows and SQL Azure is a good migration or expansion strategy.

A “hybrid” architecture merely means that part of a computing cycle is shared between two architectures. For instance, some level of computing might be done in a Windows Azure web-based application, while the data is stored locally at the organization.

Implementation:

There are multiple methods for implementing a hybrid architecture, in a spectrum from very little interaction from the local infrastructure to Windows or SQL Azure. The patterns fall into two broad schemas, and even these can be mixed.

1. Client-Centric Hybrid Patterns

In this pattern, programs are coded such that the client system sends queries or compute requests to multiple systems. The “client” in this case might be a web-based codeset actually stored on another system (which acts as a client, the user’s device serving as the presentation layer) or a compiled program. In either case, the code on the client requestor carries the burden of defining the layout of the requests.

While this pattern is often the easiest to code, it’s the most brittle. Any change in the architecture must be reflected on each client, but this can be mitigated by using a centralized system as the client such as in the web scenario.

2. System-Centric Hybrid Patterns

Another approach is to create a distributed architecture by turning on-site systems into “services” that can be called from Windows Azure using the service Bus or the Access Control Services (ACS) capabilities. Code calls from a series of in-process client application. In this pattern you move the “client” interface into the server application logic.

If you do not wish to change the application itself, you can “layer” the results of the code return using a product (such as Microsoft BizTalk) that exposes a Web Services Definition Language (WSDL) endpoint to Windows Azure using the Application Fabric. In effect, this is similar to creating a Service Oriented Architecture (SOA) environment, and has the advantage of de-coupling your computing architecture. If each system offers a “service” of the results of some software processing, the operating system or platform becomes immaterial, assuming it adheres to a service contract.

There are important considerations when you federate a system, whether to Windows or SQL Azure or any other distributed architecture. While these considerations are consistent with coding any application for distributed computing, they are especially important for a hybrid application.

Connection resiliency - Applications on-premise normally have low-latency and good connection properties, something you’re not always guaranteed in a distributed and hybrid application. Whether a centralized client or a distributed one, the code should be able to handle extended retry logic.

Authorization and Access - In a single authorization environment like a Active Directory domain, security is handled at a user-password level. In a distributed computing environment, you have more options. You can mitigate this with using The Windows Azure Application Fabric feature of ACS to make the Azure application aware of the App Fabric as an ADFS provider. However, a claims-based authentication structure is often a superior choice.

Consistency and Concurrency - When you have a Relational Database Management System (RDBMS), Consistency and Concurrency are part of the design. In a Service Architecture, you need to plan for sequential message handling and lifecycle.

Resources:

How to Build a Hybrid On-Premise/In Cloud Application: http://blogs.msdn.com/b/ignitionshowcase/archive/2010/11/09/how-to-build-a-hybrid-on-premise-in-cloud-application.aspx

General Architecture guidance: http://blogs.msdn.com/b/buckwoody/archive/2010/12/21/windows-azure-learning-plan-architecture.aspx

Windows Azure Use Case: Web Applications

BuckWoody — Mon, 14 Feb 2011 17:22:42 GMT

Description:

Many applications have a requirement to be located outside of the organization’s internal infrastructure control. For instance, the company website for a brick-and-mortar retail company may want to post not only static but interactive content to be available to their external customers, and not want the customers to have access inside the organization’s firewall.

There are also cases of pure web applications used for a great many of the internal functions of the business. This allows for remote workers, shared customer/employee workloads and data and other advantages. Some firms choose to host these web servers internally, others choose to contract out the infrastructure to an “ASP” (Application Service Provider) or an Infrastructure as a Service (IaaS) company.

In any case, the design of these applications often resembles the following:

In this design, a server (or perhaps more than one) hosts the presentation function (http or https) access to the application, and this same system may hold the computational aspects of the program. Authorization and Access is controlled programmatically, or is more open if this is a customer-facing application. Storage is either placed on the same or other servers, hosted within an RDBMS or NoSQL database, or a combination of the options, all coded into the application.

High-Availability within this scenario is often the responsibility of the architects of the application, and by purchasing more hosting resources which must be built, licensed and configured, and manually added as demand requires, although some IaaS providers have a partially automatic method to add nodes for scale-out, if the architecture of the application supports it. Disaster Recovery is the responsibility of the system architect as well.

Implementation:

In a Windows Azure Platform as a Service (PaaS) environment, many of these architectural considerations are designed into the system.

The Azure “Fabric” (not to be confused with the Azure implementation of Application Fabric - more on that in a moment) is designed to provide scalability. Compute resources can be added and removed programmatically based on any number of factors. Balancers at the request-level of the Fabric automatically route http and https requests. The fabric also provides High-Availability for storage and other components. Disaster recovery is a shared responsibility between the facilities (which have the ability to restore in case of catastrophic failure) and your code, which should build in recovery.

In a Windows Azure-based web application, you have the ability to separate out the various functions and components. Presentation can be coded for multiple platforms like smart phones, tablets and PC’s, while the computation can be a single entity shared between them. This makes the applications more resilient and more object-oriented, and lends itself to a SOA or Distributed Computing architecture.

It is true that you could code up a similar set of functionality in a traditional web-farm, but the difference here is that the components are built into the very design of the architecture. The API’s and DLL’s you call in a Windows Azure code base contains components as first-class citizens. For instance, if you need storage, it is simply called within the application as an object. Computation has multiple options and the ability to scale linearly.

You also gain another component that you would either have to write or bolt-in to a typical web-farm: the Application Fabric. This Windows Azure component provides communication between applications or even to on-premise systems. It provides authorization in either person-based or claims-based perspectives.

SQL Azure provides relational storage as another option, and can also be used or accessed from on-premise systems. It should be noted that you can use all or some of these components individually.

Resources:

Design Strategies for Scalable Active Server Applications - http://msdn.microsoft.com/en-us/library/ms972349.aspx

Physical Tiers and Deployment - http://msdn.microsoft.com/en-us/library/ee658120.aspx