Are We All Cloud Service Brokers Now? Part II

By John Dixon, Consulting Architect

In my last post, I discussed Cloud Service Brokers and some of their benefits after reading a couple of articles from Robin Meehan (Article 1 here and Article 2 here). In this post, I will break down some of Robin’s points and explain why I agree or disagree with each.

At the end of last post, I was breaking down cloud arbitrage into three areas (run-time, deployment-time, plan-time). Credit to Robin for run-time and deployment-time arbitrage. I really like those terms, and I think it illuminates the conversation. So, run-time cloud arbitrage is really science fiction right now – this is where the CSB moves running workloads around on the fly to find the best benefit for the customer. I haven’t seen any technology (yet) that does this. However, VMware does deployment-time and run-time arbitrage with VMotion and Distributed Resource Scheduling – albeit, in a single virtual datacenter, with individual VMs, and with a single policy objective to balance a cluster’s load across vSphere nodes. See Duncan Epping’s excellent write up on DRS here. Even 10 years ago, this was not possible. 15 years ago, this was certainly science fiction. Now, it’s pretty common to have DRS enabled for all of your vSphere clusters.

A few of Robin’s points…

Point 1:
“The ability to migrate IT workloads dynamically (i.e. at run-time, not at deployment time) is something I sometimes see as a capability under the ‘cloud broker’ banner, but in my view it really just doesn’t make sense – at least not at the moment.”

I agree. Run-time cloud arbitrage and workload migration ala vMotion is not possible today in cloud. Will it be possible within the next few years? Absolutely. I think it will first manifest itself in a VMware High Availability-like scenario. Again, see Duncan Epping’s fantastic deep-dive into HA. If cloud provider X drops off of the internet suddenly, then restart the resources and application at cloud provider Y (where cloud provider Y might even be your own datacenter). This is sometimes known as DR as a service, or DRaaS. And even now, there are some DRaaS solutions that are coming onto the market.

Point 2:
“The rate of innovation in the IaaS/PaaS/DaaS market is such that most of the other vendors are playing catch-up with AWS, as AWS continue to differentiate themselves from the following pack. This shows no sign of slowing down over the next couple of years – so the only way a migrated workload is going to work across multiple cloud vendors is if it only relies on the lowest common denominator functionality across the vendors, which is typically basic storage, virtualised compute and connectivity.”

Also agree, the rate of innovation in the market for cloud computing is rapid as specialization sets in at an industrial level. This also means that downward price pressures are enormous for vendors in the cloud space, even today as vendors vie for market share. As switching costs decrease (e.g., portability of applications increases), prices for IaaS will decrease even more. Now, wouldn’t you, as a customer, like to take advantage of this market behavior? Take in to consideration that CSBs aggregate providers but they also aggregate customer demand. If you believe this interpretation of the market for IaaS, then you’ll want to position yourself to take advantage of it by planning portability for your applications. A CSB can help you do this.

Point 3:
“The bottom line is that if you are going to architect your applications so they can run on any cloud service provider, then you can’t easily use any of the good bits and hence your value in migrating to a cloud solution is diminished. Not ruined, just reduced.”

Disagree. To take advantage of market behavior, customers should look to avoid using proprietary features of IaaS platforms because they compromise portability. Like we noted earlier, increased portability of applications means more flexibility to take advantage of market behavior that leads to decreasing prices.

This is where perspective on cloud becomes really important. For example, GreenPages has a customer with a great use case for commodity IaaS. They may deploy ~800 machines in a cluster at AWS for only a matter of hours to run a simulation or solve a problem. After the result is read, these machines are completely destroyed—even the data. So, it makes no difference to this customer where they do this work. AWS happens to be the convenient choice right now. Next quarter, it may be Azure, who knows? I’m absolutely certain that this customer sees more benefit in avoiding the use of propriety features (a.k.a., the “good bits” of cloud) in a cloud provider rather than using them.

What is your perspective on cloud?
• A means to improve time to market and agility
• A way to transform capex into opex
• Simply a management paradigm – you can have cloud anywhere, even internally as long as you have self-service and infinite resources
• An enabler for a new methodology like DevOps
• Simply a destination for applications

I think that a good perspective may include all of these things. Leave a comment and let me know your thoughts.

Interested in learning more? Download this free whitepaper ‘Cloud Management, Now!’