This week saw cloud related announcements from two software behemoths from the traditional era, Microsoft and Oracle. Microsoft rebooted Windows Azure making it more palatable to modern day developers and started playing nice on the interoperability game. Oracle re-announced their public cloud strategy and, in the process, tried to convince users that they should see cloud using the Oracle lens. Both these announcements are significant in their own ways and I thought I will do a quick analysis on the news.
First, the Microsoft news
Even though Microsoft’s announcement was slated for Thursday, they pre-empted Oracle’s announcement by releasing the blog post one hour before the Oracle’s webcast. So, I am tackling their news first . You can find the nitty gritty details of Microsoft’s announcements here and here. You can also read the detailed notes of the Azure event from Wade Wegner’s blog. In short, Microsoft’s announcement means the following:
- They are going down the stack to become an IaaS player. Even though I would prefer Microsoft to bypass IaaS and take a leadership position pushing PaaS, I think this move has the potential to offer the much needed competition to AWS in the public cloud market. We cannot afford to have AWS as the monopoly leader in the space and we need a strong competition. Microsoft has the necessary muscle to offer this competition. Let us see how far they can go in slowing down AWS. If their partnerships with AppFog and Apprenda are any indication, they are willing to morph into a pure play IaaS player to gain traction through their ecosystem partners. This is a smart move from Microsoft, IMO. Their support for Linux Virtual Machines is important from this angle. On a side note, I have a feeling that Microsoft and Canonical have begun their mating dance. It is just a matter of time before the acquisition happens (#plainspeculation).
- They have rebooted their PaaS strategy by rewriting their APIs to meet the modern day needs. This is what they should have done when they came out in the first place but it is better late then never. It is a right step in the right direction for Microsoft. In fact, they have completely revamped their UI to an extent I even thought it was AppFog who did the UI . Their support for building .NET, PHP and Node.js website on Azure is another good move. Now they also support distributed cache which, again, is a good move IMO. In short, Microsoft just made Azure competitive with the other smaller PaaS providers.
- I already talked about their willingness to interoperate with competitors in my previous post. I want to emphasize the importance of this move again. It will not only help Microsoft in increasing adoption for Azure, it has the potential to increase PaaS adoption, in general.
Now, the Oracle news
In short, Oracle’s news for me is a disappointment. I expected them to come up with a solid PaaS offering but what I saw was PaaS washing. On Wednesday, Oracle announced what we all knew already. They are offering their Fusion Apps on the cloud and also offering Database as a Service using their powerful database and a Java as a Service. On one hand, it shows Oracle’s realization that Cloud is here to stay and they have to embrace it in spite of Larry Ellison’s earlier dismissals. on the other hand, what they are offering does not align with the puritan definition of public cloud services where multi-tenancy plays a critical role. Oracle is making a spin that their PaaS offers the isolation provided by the virtual machines to give better security but it is not PaaS in the puritan sense. I do agree that some organizations will be comfortable with this level of isolation but my question is about the need to call this PaaS. In fact, it is akin to offering the same applications, database and runtime on top of Amazon EC2. In fact, we should give Larry Ellison credit for accepting this fact but it is not PaaS. Period. Any attempts to confuse buyers with a convenient definition of PaaS is ridiculous and short sighted. In short, hypervisor bundled with API doesn’t make it a cloud. Similarly, runtime bundled in virtual machines doesn’t make it a PaaS.
Oracle is in a great position to emerge as a PaaS leader but it can happen ONLY if they do it right and not arm twist definitions to suit their current strategy. Let us see if they become a real PaaS player by 2020 (if you followed Larry Ellison’s attack on SAP, you will get this).
Microsoft, after trying to mess with the market for a long time, is doing it right now. They understand that they don’t control today’s market like how they did in the past. This realization has lead them to do the right moves and emerge as a serious competitor in the space. Oracle is hoping to bulldoze the cloud market with their current market position but they will soon realize that it is not going to work. I just hope it is not too late for them when they have the Microsoft moment in the cloud. With their belief in open standards, they could have easily driven the cloud market away from proprietary offerings like Salesforce and others. But they are missing the boat now. Can they come back later in the game and still stay competitive? It is possible but I am not sure if they will ever want to do it right.
Disclosure: AppFog and Apprenda are sponsors of DeployCon 2012
- Oracle and the Vision Thing (enterpriseirregulars.com)
- Oracle’s Behavior Undercuts Its Own Cloud Accomplishments (Enterprise System Spectator)