Using the Amazon cloud for Dev/Test in 5 easy steps

December 6, 2011 Comments Off on Using the Amazon cloud for Dev/Test in 5 easy steps

Recently, I explained why the software development and testing process (Dev/Test) is a fantastic use case for cloud computing.  At first glance, getting started might be a bit daunting, but it’s really quite simple.

  1. Determine whether you want to use the cloud for development, testing, or both. If you’re building an Internet-based solution (with or without a user interface), the cloud is particularly adept at letting you quickly set up distributed load tests. These tests – which can be provisioned globally – provide much more accurate metrics about the true response times that your users will experience than if you tried to run load tests from a single location, usually within your firewall!
  2. Identify the cloud that you want to use. I’ll describe the major cloud computing platforms in a series of upcoming posts. For now, the Amazon cloud – specifically Elastic Compute Cloud – is a great place to start, especially for Dev/Test.
  3. Assuming you’ve chosen Amazon, set up a single instance of a virtual machine either on your computer, or select from one of the hundreds of pre-configured Amazon Machine Images. On the other hand, if you’ve created your own virtual machine you can then upload it to Amazon.
  4. Once you’ve set up your base virtual machine and it’s running on Amazon, add all of the ancillary software and technology necessary to complete your Dev/Test environment. These can include:
    • Software development tools
    • Database engines
    • Application servers
    • Web servers
    • Security software
  1. When your virtual machine is fully and properly staged, you can easily clone it as many times as you need.

Going forward, if your Dev/Test cloud experience is positive, consider implementing specialized software that automates many of the manual steps that were necessary to create the cloud-based environment. I’ll describe this specialized software in a future posting.

Technical marketing mistake #2: Failing to consider ROI

December 4, 2011 § 3 Comments

At the start of this series, I listed some of the most common technical marketing oversights that we see at Think88. In this installment, I describe why it’s so important to consider return-on-investment before developing any marketing collateral.

In far too many companies, leaders and rank-and-file alike are convinced that the sales department exists to bring in money, and the marketing department exists to gleefully spend it. This perception is a gross oversimplification, and is largely unwarranted in most cases. However, there are certain situations where this observation is justified. Marketing collateral is one of those scenarios, but there’s plenty of blame to go around:

  •  Vocal salespeople who clamor for something – anything! to give to their prospects.
  • Marketing executives eager to satisfy these demands.

Both sales and marketing are far too reactive in these cases. This is very unfortunate, because creating this collateral can be expensive, time-consuming, and starve other (often more qualified) initiatives of necessary resources.

Instead of making knee-jerk investments in new collateral, it’s much wiser to first subject any potential materials to a rigorous ROI assessment. Of course, ROI should also be a factor in the marketing content roadmap that I described in an earlier post. I’ve found that the most effective marketing collateral is produced after analyzing weak points in the sales cycle, and then creating materials to plug those gaps. Frankly, there is no substitute for ‘riding along’ with salespeople, or experiencing the sales cycle as a prospect would. With proper research, it should be possible to quantify – or at least estimate – the financial impact of any marketing collateral investment.

What does one of these ROI calculations look like? I’ll provide a real-world example in a future posting.

Are you at risk from escalating cloud storage costs?

December 2, 2011 Comments Off on Are you at risk from escalating cloud storage costs?

The damage wrought by the recent Thai floods has caused a significant run-up in disk drive prices. I described this situation in an earlier post.

While these increases have primarily been confined to the consumer marketplace, there’s a good chance that they will soon hit business users, too. Even though some of these wrecked factories are coming back online (or being moved to other locations), rising energy costs, soaring demand, and a shrinking number of disk drive suppliers can only exacerbate the impact of these potential price increases. Unfortunately, storage price increases will be counter to the trend – largely driven by Amazon – of steadily decreasing costs for commoditized storage.

Naturally, cloud is a voracious consumer of data, so just about every cloud solution may be impacted. But businesses are particularly at risk in two key cloud scenarios:

1. Big Data. More data than ever is being captured, managed, and analyzed. The cloud has been a natural fit in this situation: in a preponderance of cases, it’s more cost-effective and efficient than building internal infrastructure. An entire ecosystem is growing around cloud-based Big Data, with cloud-based integration and analytics just two examples of these new storage-hungry applications.

2. Cloud-based backup. In the past few years, enterprises have been adopting cloud-connected storage in lieu of onsite backup. There are many compelling reasons for this transition, such as strong vendor SLAs, economies of scale, and reduced labor costs. If you’re curious about this topic, take a look at the paper for i365 we wrote at Think88 .

Both of these scenarios consume enormous amounts of storage, and are highly sensitive to sudden, dramatic increases in storage costs such as we may experience.

Whether price increases are sudden or gradual, what should you do? I’ll explore that topic in an upcoming posting.

Where Am I?

You are currently viewing the archives for December, 2011 at rdschneider.

%d bloggers like this: