Skip to content

RH Partner summit: full-speed, all in the same direction toward cloud

During my career, I have seen a lot of companies, especially multi-national companies, being disrupted by internal wars, arguments between departments, jealousy between projects and much more that you can imagine. What’s really impressive after one day assisting at the Red Hat partner summit conferences, is the clear vision, direction and alignment of the whole Red Hat products and projects in the same and unique direction: enterprise, data-centre and cloud.

From the OS to the middleware

RH is going full speed to the cloud and step-by-step aligning all the key components on each layer from the OS to the middleware. From the OS we can find the Red Hat Enterprise Virtualization (RHEV) which is basically composed by KVM, the hypervisor, and by a management console. Today RHEV is mainly focus on virtualization portability with the OVF import and the V2V support for VM conversion, highly manageable virtual cluster through RHN Satellite, integration with reporting tools and finally an Ecosystem of app provider-ISV-HW manufacturers  certified RHEV. The current strategy is then achieving a complete virtualization framework and securing the applications on top. But for next year the roadmap clearly takes a clear direction:

  • SLA Support: defining resource requirements for an image, network expectation, etc. and the ability to achieve this SLA at runtime.
  • Virtual M-Physical M scheduler: an hybrid model for scheduling job either on the Virtual machines or physical machines.
  • Enhancement of the security manager and isolation of VI.

 

This sounds very interesting for a cloud infrastructure…

On the middleware side we can see JBoss ON [2] really improved for managing a JBoss cluster, API for server provisioning. I have even seen a cool demo by Graham Gear on which a custom script automatically called yum for provisioning a set of machines with JBoss AS, deploying application and setting up a buddy cluster configuration. This really looks like interesting in a cloud …

In addition,  the Messaging framework such as HornetMQ [3] can be used for connecting all node instances and providing high speed communication between AS.

On top of JBoss we have the SOA platform

As soon as we have JBoss in the house, we can build the entire integration and SOA platform provided by RH. I will not say that the SOA-P has been designed for the cloud, but the first customers of the cloud are enterprises and one of their key requirements is clearly the integration. Actually the first one is to use the usual software such as SAP, Tivoli of this world, which is covered by the RH Ecosystem strategy and therefore the second is the integration between those softs and internal applications, that’s where the SOA-P comes. As a result, just like the RHEV, JBoss SOA-P can be used on physical traditional machines, but also on the cloud.

Not yet a cloud platform

However, the developer still needs to design himself the architecture : where will be located my data, do I need a distributed cache such as JBoss Cache or INFINISPAN, what will be my replication policies, how do I manage the load balancing between the Amazon LB and the mod_cluster, how Amazon will be aware of my buddy configuration, etc. As a result the RH’s offer today, looks more an Infrastructure offer for the cloud than a platform offer for the cloud. But they are working hard on this topic to provide such kind of platform for 2011.

This is in-line with the strategy

As explained by Scott Crenshaw in his keynote, the RH strategy for getting customer into the cloud is defined by 3 phases:

  1. Virtualization: virtualize environments at customer premises or in data-centres
  2. Private cloud: building private cloud for the customers
  3. Hybrid cloud: integrating public cloud when needed, it’s the elasticity on top of elastic architecture

 

The virtualization is ready, the private cloud is currently an infrastructure cloud and the hybrid cloud is coming in 2011. RH has clearly defined the direction to take and is aligning all its products in this double approach: for the cloud and not for the cloud.

References

[1] The RHEV web site, http://www.redhat.com/virtualization/rhev/

[2] The JBoss ON, http://www.jboss.com/products/jbosson/

[3] Hornet MQ, www.jboss.org/hornetq

[4] The SOA-P, http://www.jboss.com/products/platforms/soa/


 

Releated Posts

Calibrate to Interpret

Trustworthy machine learning is driving a large number of the ML community works in order to improve ML acceptance and adoption. In this paper, we show a first link between uncertainty and explainability, by studying the relation between calibration and interpretation.
Read More

Mass Estimation of Planck Galaxy Clusters using Deep Learning

Galaxy cluster masses can be inferred indirectly using measurements from X-ray band, Sunyaev-Zeldovich (SZ) effect signal or optical observations. Unfortunately, all of them are affected by some bias. Alternatively, we provide an independent estimation of the cluster masses from the Planck PSZ2 catalogue of galaxy clusters using a machine-learning method.
Read More