mardi, juillet 29, 2014

Amazon AWS grows slowly : not (yet) appropriate for mission-critical applications

In his post on NYT blogs, Quentin Hardy reports about slow growth of Amazon AWS IaaS services (S3, EC2, etc.) : "only" +37%  (1.2 billion $ revenues for the quarter)  when compared to a year ago. It was far above 50% in the past !

Of course, you can't grow forever at this huge speed but something else is happening.
I would agree with the statement by Chris Gaun in this post saying the public cloud is not ready for the enterprise (or vice-versa).

The "impedance mismatch" between public cloud offerings (Amazon AWS, Microsoft Azure, Google Cloud) and corporation requirements is big : SLAs are not stringent enough, security is not guaranteed, etc. on the cloud providers side. On the corporation side, it is hard to express the right needs in terms of standardized cloud services when it comes to move an internal application (usually made of various satellite subsystems) usually setup over many years.

That is cleary why many corporations prefer to learn and test the waters with a private cloud behind their firewall : I already reported why and how CIOs prefer to see their mission-critical applications (usually running on mainframes) migrated to an internal CloudStack or OpenStack system for the Java and web UI that we generate at Eranea to replace their Cobol and 3270.

It is the safest path to learn and experience daily cloud computing at large scale for core business applications. At the same time, they get ready for a much simpler move to AWS, Azure & co : moving already virtualized instances from a private cloud to a shared infrastructure is much simpler than the direct jump from a mainframe to a public cloud.

So, CIOs lose no time by doing the intermediate step of private cloud : they just prepare for the future (=hybrid cloud) by implementing the leading edge for mission-critical application (= private cloud)

Source: blog Media and Tech (par didier durand)

samedi, juillet 26, 2014

Transitioning legacy to the cloud : feeling of inability rather than missing willingness ! (2014 Future of Cloud Survey)

For those interested in future of cloud computing, have detailed look at the 2014 Future of Cloud Survey by Northern Bridge (shared below from SlideShare): a very comprehensive analysis of current market trends for all aspects of the cloud. With 1,350+ respondents from all kinds of companies, organizations, etc., it clearly is the largest survey in this area.

Let's comment some aspects of this survey from Eranea's perspective :
  1. The quote of Mike Schutz from Microsoft (page 7) is definitely true : “what comes to me loud and clear from this survery is that even more business are no longer thinking 'why cloud' but will focus the next 12-24 months figuring out 'how do I execute a long-term cloud strategy?'
  2. Some functions of corporate IT have won the cloud battle (page 22) : web presence, communications, disaster recovery.
  3. The laggards are in the back-office (page 24): transaction processing and data center consolidation
  4. Security is still a big concern (page 55) : it remains the biggest inhibitor to cloud computing for 49% of respondents
Our current and frequent meetings with prospects and clients (throughout various countries in Europe) who wants to leave transform their IT legacy (Cobol applications, mainframe systems) bring the same feedback:

  1. CIOs no longer question the interest of the cloud for their company. Its advantages (cost savings, flexibility & agility, scalability, etc.) are perfectly understood. But, for their legacy systems, they have a slightly harder question than the one proposed by Mike Schutz: it's not “how do I execute a long-term cloud strategy”, it is more “how can I execute a long-term strategy cloud strategy” (for my legacy). Ability is their concern.
  2. Of course, they already migrated or are about to migrate the “cloud-native” functions of their global IT system: cloud computing is born with the Internet (remember those service providers named (web) hosting companies ?). So, there is no better place to for web presence and communications services (email, etc.) than the cloud. Then, regarding disaster recovery, its cloud version is so cost-efficient that it is for many companies an enabler: finally, they can afford a solid recovery architecture for their mission-critical applications !
  3. Yes, the laggards are in the back-office, but for a good reason : CIOs have the deep feeling of being stuck (trapped ?) with their big irons and the associated millions of lines of Cobol. No way out !
  4. Yes, security is a big concern regarding shared cloud infrastructures: wouldn't you be worried as a bank or financial service company to see your IT under deep scrutiny (spying ?) if data in your core banking system run by a cloud provider could “leak” at any time ? Remember the huge story of 2013 involving NSA, PRISM, etc. ? Such scary megabuzz leaves imprints in the brains of most pioneer CIOs ...

At Eranea, we try to address those concerns in the following way:
  • We address the feeling of inability to migrate legacy systems by detailing how we can migrate their millions of lines of Cobol running on their mainframe in a very safe manner. Our process of transcoding to Java and web is 100% automated. The code that we generate is strictly iso-functional allowing very progressive and riskless migrations. So, we demonstrate that the transformation of Cobol to a Java + web application able run in  a cluster of cloud Linux instances is, for Eranea, an industrial and progressive process (i.e. repeatable at will thanks to appropriate technology and tooling) and NOT a 1-shot extraordinary achievement run in big-bang mode.
  • We propose to use private cloud initially: it keeps everything internal to the company. So, the security concern is blown away. Secondly, it is an “acceptable” first step toward cloud computing as everything stays under full control of internal IT organization. Private cloud solutions like CloudStack or OpenStack are already productive at very high-scale in very large corporations (Disney, Wells Fargo Bank, etc.)  or organizations (CERN, NASA, etc.) : the services that those private cloud propose allow to reach architecture matching the most stringent requirements coming from the continuation of mainframe SLAs.
Private cloud is the right way to start the “cloudification” of your legacy: it allows you to learn this still-emerging technology “behind the curtain.” So, you avoid the public exposure (media buzz, hackers, etc.) around your initial missteps on mission-critical applications.

When your corporation feels expert in subtleties of cloud computing via learning on a private cloud,, it's probably time to go for the shared version of cloud services, be it named hybrid cloud, intercloud, multicloud PaaS, Iaas, etc.

Isn't it a wise step anyway if you refer to the old adage : “delegation without understanding is abdication” ?

So, to come back to the “Future of Cloud ” 2014 survey, the figures around legacy strongly reflects the current market trends.

And, from our perspective,  they are motivated by a feeling of inability by CIOs: “I am locked up can't get there for my legacy.” On the other side, as their willingness is strongly present and as smooth and secure migration solutions like the one we continuously perfect at Eranea exist, it is not so hard to convince them that the world famous “YES, WE CAN !” also applies to the cloud strategy of their legacy system.

Source: blog Media and Tech (par didier durand)