Welcome!

From the Author of The Agile Architecture Revolution

Jason Bloomberg

Subscribe to Jason Bloomberg: eMailAlertsEmail Alerts
Get Jason Bloomberg via: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


Related Topics: Cloud Computing, SOA Best Practices Digest, Microservices Journal

Cloud Computing: Article

Beware Fake Clouds

There’s still plenty of confusion over the what Cloud is and how to achieve value with it

First the good news: Cloud Computing is real, it’s here to stay, we’re doing a reasonably good job of defining it, and it’s already providing significant business value to countless organizations around the globe. Now the bad news: there’s still plenty of confusion over the what Cloud is and how to achieve value with it.

Much of this confusion, as you might expect, comes from software and hardware vendors. After all, they’ve all had to scramble for a new Cloud value proposition once they realized that Cloud Computing would eventually doom the old way of selling their gear. For most such vendors, their Cloud strategies are works in progress. Yes, they may have a bona fide Cloud offering, but if you look more closely, the Cloud benefits you’re expecting may not yet be available. Our advice? Caveat emptor.

Getting the Elasticity You Require
In spite of the title of this ZapFlash, the distinction between a “real” or “fake” Cloud isn’t particularly useful, since it could be argued that any subscription-based Web site is a simple example of SaaS. What really matters is the value proposition. If all you want is a pay-as-you-go subscription model for something you access over the Web, then virtually any vendor’s purported SaaS offering may qualify. However, a subscription model doesn’t guarantee elasticity or automated recovery from failure, two essential Cloud characteristics. If you don’t care about these characteristics, then fine. But don’t be fooled. A vendor may say their offering is Cloud-based, suggesting they have an elastic offering even if they don’t.

It’s also important to understand the different types of elasticity. Even if a vendor says their offering is elastic, you may need to dig further. They may simply be referring to the elasticity of their virtualization layer. An IaaS provider might offer you, say, a virtual machine (VM) with a gigabyte of RAM, with the promise that if you need two gigabytes, you’ll get it automatically, and only pay for it while you’re using it. Yes, this is a form of elasticity, but it has limits. After all, your VM is rubbing elbows with other VMs on some physical server with physical memory somewhere, and there’s only so much RAM to go around. Allotting you more might even mean borrowing it from someone else’s VM.

However, you may be looking for the unlimited type of elasticity that gives Clouds the illusion of infinite capacity—in other words, the elasticity that makes Clouds cloudy. For this type of elasticity, what we might call Cloud elasticity to distinguish it from the limited form in the paragraph above, the Cloud provider must be able to provision and deprovision additional instances quickly and automatically, where “instances” might refer to VMs, storage, queues, databases, or whatever resources you’re interested in acquiring from the provider. IaaS vendors find this kind of horizontal elasticity relatively straightforward, since it’s up to you how you’re going to use your new instances. But for PaaS and SaaS vendors, Cloud elasticity can be unexpectedly problematic.

For example, take a look at the Oracle Database Cloud. This offering essentially takes the enterprise workhorse Oracle Database 11g and places it into a virtualized environment—what Oracle refers to as a PaaS offering in a Private Cloud. The architectural emphasis, however, is on database consolidation, not horizontal elasticity. The problem is that the Oracle Database is inherently partition intolerant, because it guarantees availability and immediate consistency. Their offering may very well meet your needs, but don’t expect it to offer Cloud elasticity.

It’s also important to question your SaaS or PaaS provider about multitenancy. As we discussed in an earlier ZapFlash, there are several different flavors of multitenancy, and they support different value propositions. If a vendor has a traditional app and they want to bring it to the Cloud market quickly, they will typically offer a shared-hardware or shared-OS megatenancy model. With megatenancy, the vendor simply installs the same software they sell commercially on top of virtualized infrastructure, one instance per customer, and then offers customers pay-as-you-go access. If that sounds like a hosted provider model as opposed to a true Cloud model, you’re on the right track—although you may only require pay-as-you-go pricing via a hosted provider model, so the distinction may be moot. But if you’re looking for the elasticity, collaboration capabilities, and coordinated, transparent updates of a SaaS offering with shared-table multitenancy, then be sure your provider truly offers them.

Tough Questions to Ask
There’s more to getting what you pay for in the Cloud than ensuring the elasticity and multitenancy on offer meet your needs. Here are some tough questions you can ask to separate the wheat from the chaff:

  • Did the SaaS provider simply cross out the word “Web” on their marketing and replace it with “Cloud”? Yes, this might be all you require, but chances are you’re looking for something more. Remember the dot.com days where mundane companies would stick the word “Web” in their marketing and automatically become a dot.com player? Well, now the spinmeisters are doing it again. Accessing software with a browser over the Internet doesn’t make it “Cloud.” That’s what we used to call the Web.
  • Will moving to the Cloud really save me money? Saving money may or may not be your reason for moving to the Cloud, but for many organizations, it’s their primary business driver. However, if your capacity requirements are relatively stable—that is, elasticity isn’t particularly important to you—then IaaS in particular may actually be more expensive than just leaving your apps where they are. Make sure you crunch the numbers before taking the plunge.
  • Do you like the idea of SaaS, but no existing SaaS offering is quite right, so you’re thinking about hiring someone to build you a “custom” SaaS solution? If so, you’re almost certainly on the wrong track. The whole point to SaaS is you’re leaving the software development as well as hosting work to someone else who can make money from many customers, thus lowering the cost for all of them. Paying someone to build a bespoke solution defeats the whole purpose. It’s really not SaaS at all, even if you access it over the Web.
  • Does your PaaS provider’s platform give you the APIs you require? For PaaS providers who’ve built their platforms from scratch to run in the Cloud, this is a silly question. Take the Facebook app platform, for example. The whole point of running your app on Facebook is to interface with the core Facebook app, so of course they provide the APIs you need to do so. But what if your PaaS provider took some old middleware product, say an ESB, stuck it in the Cloud, and called it PaaS? You’ll be lucky if the APIs you get simply reflect the fact they’re running in the Cloud at all, let alone offer you specialized capabilities unique to the environment.
  • Just how automated is a public Cloud provider’s automated provisioning and configuration? Elasticity doesn’t just require dynamic provisioning and deprovisioning, it requires automated provisioning and deprovisioning. If provisioning a VM means sending a work order to a sysadmin who’ll get back to you in a few days, it’s not Cloud at all.
  • Similarly, are you considering investing in a Private Cloud, but your vendor can’t provide fully automated configuration and provisioning tools? If so, they’re pulling the wool over your eyes. How do you expect to handle configuration and provisioning, by hiring a room full of monkeys pounding on keyboards all day? Puhleeze. We have a phrase for a Private Cloud without automated configuration and provisioning. We call it a traditional data center.

The ZapThink Take
Fortunately, ZapThink is here to help. I’m speaking at a number of conferences over the next few months, and many of them have vendor-heavy agendas. To balance all that vendor spin, my talks always focus on how to achieve real business value by understanding what the vendors can—and cannot—provide. As an added aid, feel free to print this ZapFlash and bring it along. It’ll help you ask the right questions when you face that smiling sales rep at the exhibit booth. I hope to see you in London this Friday, Enterprise Data World in Atlanta on May 2 (discount code SPEAKER429) , Dallas on May 22, Kuala Lumpur on May 28 – 29, or Cloud Expo in New York June 11 – 14 (Golden Pass discount code zapthinkdiscount). I want to see you with this cheat sheet in hand!

Photo credit: Jayel Aheram

More Stories By Jason Bloomberg

Jason Bloomberg is the leading expert on architecting agility for the enterprise. As president of Intellyx, Mr. Bloomberg brings his years of thought leadership in the areas of Cloud Computing, Enterprise Architecture, and Service-Oriented Architecture to a global clientele of business executives, architects, software vendors, and Cloud service providers looking to achieve technology-enabled business agility across their organizations and for their customers. His latest book, The Agile Architecture Revolution (John Wiley & Sons, 2013), sets the stage for Mr. Bloomberg’s groundbreaking Agile Architecture vision.

Mr. Bloomberg is perhaps best known for his twelve years at ZapThink, where he created and delivered the Licensed ZapThink Architect (LZA) SOA course and associated credential, certifying over 1,700 professionals worldwide. He is one of the original Managing Partners of ZapThink LLC, the leading SOA advisory and analysis firm, which was acquired by Dovel Technologies in 2011. He now runs the successor to the LZA program, the Bloomberg Agile Architecture Course, around the world.

Mr. Bloomberg is a frequent conference speaker and prolific writer. He has published over 500 articles, spoken at over 300 conferences, Webinars, and other events, and has been quoted in the press over 1,400 times as the leading expert on agile approaches to architecture in the enterprise.

Mr. Bloomberg’s previous book, Service Orient or Be Doomed! How Service Orientation Will Change Your Business (John Wiley & Sons, 2006, coauthored with Ron Schmelzer), is recognized as the leading business book on Service Orientation. He also co-authored the books XML and Web Services Unleashed (SAMS Publishing, 2002), and Web Page Scripting Techniques (Hayden Books, 1996).

Prior to ZapThink, Mr. Bloomberg built a diverse background in eBusiness technology management and industry analysis, including serving as a senior analyst in IDC’s eBusiness Advisory group, as well as holding eBusiness management positions at USWeb/CKS (later marchFIRST) and WaveBend Solutions (now Hitachi Consulting).