Category Archives: Openstack

Hybrid World… Why IBM buying RedHat makes sense!

As Red October came to a close…at a time when US Tech stocks were taking their biggest battering in a long time the news came out over the weekend that IBM had acquired RedHat for 34 billion dollars! This seems to have taken the tech world by surprise…the all-cash deal represents a massive 63% premium on the previous close of RedHat’s stock price…all in all it seems ludicrous.

Most people that I’ve talked to about it and from reading comments on social media and blog sites suggests that the deal is horrible for the industry…but I’ve felt this is more a reaction to IBM than anything. IBM has a reputation as swallowing up companies whole and spitting them out the other side of the merger process a shell of what they once were. There has also been a lot of empathy for the employees of RedHat, especially from ex-IBM employees who have experience inside the Big Blue machine.

I’m no expert on M&A and I don’t pretend to understand the mechanics behind the deal and what is involved…but when I look at what RedHat has in its stable, I can see why IBM have made such an aggressive play for them. On the surface it seems like IBM are in trouble with their stock price and market capitalization falling nearly 20% this year and more than 30% in the last five years…they had to make a big move!

IBM’s previous 2013 acquisition of SoftLayer (for a measly 2 billion USD) helped them remain competitive in the Infrastructure as a Service space and if you believe the stories, have done very well out of integrating the SoftLayer platform into what was BlueMix, and is now IBM Cloud. This 2013 Forbes article on the acquisition sheds some light as to why this RedHat acquisition makes sense and is true to form for IBM.

IBM sees the shift of big companies moving to the cloud as a 20-year trend…

That was five years ago…and since then a lot has happened in the Cloud world. Hybrid cloud is now the accepted route to market with a mix of on-premises, IaaS and PaaS hosted and hyper-scale public cloud services being the norm. There is no one cloud to rule them all! And even though AWS and Azure continue to dominate and be front of mind there is still a lot of choice out there when it comes to how companies want to consume their cloud services.

Looking at RedHat’s stable and taking away the obvious Linux distro’s that are both enterprise and open sources the real sweet spot of the deal lies in RedHat’s products that contribute to hybrid cloud.

I’ve heard a lot more noise of late about RedHat OpenStack becoming the platform of choice as companies look to transform away from more traditional VMware/Hyper-V based platforms. RedHat OpenShift is also being considered as an enterprise ready platform for containerization of workloads. Some sectors of the industry (Government and Universities) have already decided on their move to platforms that are backed by RedHat…the one thing I would comment here is that there was an upside to that that might now be clouded by IBM being in the mix.

Rounding out the stable, RedHat have a Cloud Suite which encompasses most of the products listed above. CloudForms for Infrastructure as Code, with Ansible for orchestration…together with RedHat Virtualization together with OpenStack and OpenShift..it’s a decent preposition!

Put all that together with the current services of IBM Cloud and you start to have a compelling portfolio covering almost all desired aspects of hybrid and multi cloud service offerings. If the acquisition of SoftLayer was the start of a 20 year trend then IBM are trying to keep themselves positioned ahead of the curve and very much in step with the next evolution of that trend. That isn’t to say that they are not playing catchup with the likes of VMware, Microsoft, Amazon, Google and alike, but I truly believe that if they don’t butcher this deal they will come out a lot stronger and more importantly offer valid completion in the market…that can only be a good thing!

As for what it means for RedHat itself, their employees and culture…that I don’t know.

References:

https://www.redhat.com/en/about/press-releases/ibm-acquire-red-hat-completely-changing-cloud-landscape-and-becoming-world%E2%80%99s-1-hybrid-cloud-provider

IBM sees the shift of big companies moving to the cloud as a 20-year trend

Ninefold: Going Head to Head with AWS and Using Opensource is Risky in Cloud Land

Today Ninefold (an Australian based IaaS and PaaS) provider announced that they where closing their doors an would be migrating their clients to their parent companies (Macquarie Telecom) cloud services. And while this hasn’t come as a surprise to me…having closely watched Ninefold from it’s beta days through to it’s shutdown it does highlight a couple of key points about the current state of play in the public cloud in Australia and also around the world.

As a disclaimer…this post and the view points given are totally my own and I don’t pretend to understand the specific business decisions as to why Ninefold decided to shut up doors apart from what was written in the press today around operational expenditure challenges of upgrading the existing platform.

“After an evaluation of the underlying technical platform, much consideration and deep reflection, we have decided not to embark on this journey,” the company said on Monday.

However, rather than have people simply assume that the IaaS/Cloud game is too hard given the dominance of AWS, Azure and Google I thought i’d write down some thoughts on why choosing the right underlying platform is key to any Clouds success…especially when looking to compete with the big players.

Platform Reliability:

Ninefold had some significant outages in their early days…and when I say significant, I mean significant…we are talking days to weeks where customers couldn’t interact or power on VM instances to go along with other outages all of which I was led to believe due to their adoption of CloudStack and Xen Server as their hypervisor. At the time I was investigating a number of Cloud Management Platforms and CloudStack (at the time) had some horror bugs which ruled out any plans to go with that platform at the time…I remember thinking how much prettier the interface was compared to the just released vCloud Director but the list of show stopping bugs at the time was enough to put me off proceeding.

Platform Choice:

CloudStack was eventually bought by Citrix and then given to the Apache Foundation where is currently resides but for Ninefold the damage to their initial reputation as a IaaS provider for mine did not survive these initial outages and throughout it’s history attempted to transform firstly into a Ruby On Rails platform and more recently looked to jump on the containers bandwagon as well as trying to specialize in Storage as a Service.

This to me highlights a fairly well known belief in the industry that going Opensource may be cheap in the short term but is going to come back and bite you in some form later down the track. The fact that the statement on their closure was mainly focused around the apparent cost of upgrading their platform (assuming a move to Openstack or some other *stack based CMP) highlights the fact that going with more supported stacks such as VMware ESXi with vCloud Director or even Microsoft Hyper-V with Azure is a safer bet long term as their are more direct upgrade paths version to version and there is also official support when upgrading.

Competing against AWS Head On:

http://www.itnews.com.au/news/sydney-based-cloud-provides-price-challenge-247576

Macquarie Telecom subsidiary Ninefold launches next week, promising a Sydney-based public cloud computing service with an interface as seamless as those of Amazon’s EC2 or Microsoft’s Azure.

Ninefold from the early days touted themselves as the Public Cloud alternative and their initial play was to attract Linux based workloads to their platform and offer very very cheap pricing when compared to the other IaaS providers at the time…they where also local in Australia before the likes of AWS and Azure set up shop locally.

I’ve talked previously about what Cloud Service Providers should be offering when it comes to competing against the big public cloud players…offering a similar but smaller slice of the services offered targeting their bread and butter will not work long term. Cloud Providers need to add value to attract a different kind of client base to that of AWS and Azure…there is a large pie out there to be had and I don’t believe we will be in a total duopoly situation for Cloud services short to medium term but Cloud Providers need to stop focusing on price, so much as quality of their products and services.

Final Thoughts:

Ninefold obviously believed that they couldn’t compete on the value of their existing product set and due to their initial choice of platform felt that upgrading to one that did allow some differentiation in the marketplace compared to the big public cloud players was not a viable option moving forward…hence why their existing clients will be absorbed into a platform that does run a best of breed stack and one that doesn’t try to complete head to head with AWS…at least from the outside.

“Those tier two ISPs and managed services outfits standing up wannabe AWS clones cobbled together out of bits of Xen, OpenStack and cable ties?”
Roadkill.
As the industry matures, smaller local players will find they can’t make it pay and go away. The survivors will move into roles as resellers and managed services providers who make public cloud easier for those who don’t like to get hands on with the big boys. This is happening already. By 2015 we’ll see exits from the cloud caper.“

http://www.zdnet.com/article/ninefold-to-shut-operations/

http://www.itnews.com.au/news/ninefold-to-shut-down-411312?utm_source=twitter&utm_medium=social&utm_campaign=itnews_autopost

http://www.crn.com.au/News/411313,macquarie-telecoms-ninefold-closing-down.aspx?utm_source=twitter&utm_medium=social&utm_campaign=crn_autopost

http://www.theregister.co.uk/2013/12/19/australias_new_year_tech_headlines_for_2015/

Platform9 – GA of Managed OpenStack for VMware vSphere

Today, Platform9 have announced the General Availability of their Managed OpenStack for VMware vSphere environments which adds to Platform9’s ability to manage KVM based environments with the SaaS based OpenStack Platform. They have also announced a new round of funding:

Today Platform9 also announced that it has raised $10 million in Series B funding from Menlo and Redpoint Ventures

Both announcements come in the lead up to VMworld 2015 ensuring that there will lots of interest in the Platform9 team to build on the success they had last year. It was at VMworld last year that I was first introduced to the product and I’ve been testing both the KVM and more recently the VMware Platforms over the last couple of months. The VMware Management has particular interest to me and I can see massive benefits in extending an existing VMware Based environment to be managed by Platform9’s OpenStack.

I’ve put together an initial install and configure video showing how easy it is to get Platform9 managing an existing vCenter.

Key Benefits for VMware Customers
  • Self-Service Automation. Empowering developers is a key CIO priority in the enterprise. With multi-tenancy, quotas, leases and policy controls on resource configurations for workloads, Platform9 Managed OpenStack makes it easy for IT to implement self-service for developers and accelerate their workflows.
  • 100% Interoperability with VMware vSphere. Customers can perform any choice of operations via Platform9 Managed OpenStack or via VMware vSphere: therefore, all existing processes, vSphere based API automation and 3rd party products are compatible and supported.
  • Orchestration using Open, Industry Standard APIs. OpenStack APIs are supported by a wide range of automation tools, development libraries, storage and network systems and application-level orchestration frameworks. By orchestrating their infrastructure using OpenStack APIs, organizations not only reduce reliance on proprietary ecosystems, but also accelerate API driven infrastructure automation and leverage existing open source integrations.
  • Single Pane Across Virtualization Platforms. Platform9 now supports both KVM and VMware vSphere virtualization and Docker support is expected in future. Customers can now decouple their private cloud platform from the virtualization platform, and retain the flexibility to deploy new platforms over time while maintaining a consistent workflow and management interface.
  • Compatibility with all vSphere Compatible Datastores and Networks. Customers can onboard Platform9 with any storage and network implementation that is compatible with VMware vSphere. The solution is fully compatible with both legacy VLAN based networks as well as software-defined-networks.
  • Production Grade SLA for OpenStack. Platform9 takes care of monitoring, troubleshooting and upgrading the OpenStack framework, so customers can rely on a production grade SLA and focus on innovation enabled by the OpenStack cloud platform.

For an introduction to Platform9 see my previous blog post here:

Well done to the Platform9 team for the GA and the Series B!

http://www2.marketwire.com/mw/release_html_b1?release_id=1212828
http://www2.marketwire.com/mw/release_html_b1?release_id=1212827

Platform9 Introduction

I came across Platform9 while wandering the back halls of the VMWorld Solutions Exchange last year in San Francisco…as a fan of the movie District 9 I was drawn to the name without really knowing anything about the tech being shown. After a brief chat with the booth staff going over product I thought to myself that there was potential in a SaaS based Cloud Management/Provisioning Platform…another side of me thought it also threatens part of the work I do in designing and managing Service Provider based Cloud Platforms like vCloud Director.

A few months later Platfrom9 has launched with fresh rounds of VC funding and is ready to go prime time. I spent some time today with Sirish Raghuram Who is the CEO and Co-founder…he also has a very interesting pedigree having previously been with VMware since 2002 working on products like Workstation, SRM and vCloud Director…in fact he was part of the team responsible for bringing maturity to later versions of vCD.

Sirish took me through the basics of what Platform9 offers and how it easily plugs into on-premises Hypervisior resources focusing on KVM with BETA support for ESXi…effectively what you get with Platform9 is your own version of the Openstack Platform and your own Management Portal to control compute, storage and network resources. You then to carve up those resources for use by departments or clients. Like vCloud Director, it abstracts management and provides a mechanism to consume IaaS resources.

What I like about this solution is that its easy…Implementing Openstack is not! …Especially if you choose to deploy and manage it yourself. The time savings along with budget savings compared with deploying and maintaining OpenStack in-house make this an attractive option. It also potentially fills a gap in the market that’s been vacated by VMware’s decision to pull vCD from Enterprise and replace it with the more complicated (dare I say bloated) vRealize Automation. To clarify that comment, think about a situation where a company has a small vCenter Instance with two or three hosts…vRA isn’t the best fit if the company wants to explore Private Cloud…Platform9 fits in nicely and can be retrofitted with ease.

For Service Providers if offers an opportunity to experience Openstack and take advantage of its KVM strengths…dual stacks become a more plausible option and with access to refined APIs it makes it easy for SPs with existing Account/Control Panels to integrate and add to existing offerings. There is also a lot of interest in it’s potential for management of Docker…which is all the rage these days.

For me, this isn’t going to replace vCloud Director SP for vSphere Platforms short to medium term…there are a lot of holes in what Openstack does compared to vCloud Director but there is a future here and there are many use cases where Platform9 makes sense.