News that HP has launched its first public cloud services and will enter public beta on May 10 provoked my interest from many angles. As a former CTO of a major HP business, I had observed (and sometimes participated in) many attempts within the company to construct a cogent approach to cloud services. Despite much hype and bluster and a fair amount of wasted (and expensive) effort, nothing much had come from these attempts. HP just couldn’t make its mind up whether it would be a provider of hardware to cloud vendors or plunge into the maelstrom itself.
However, looking at what’s now on the table, I think that there are a couple of reasons why HP has a shot at being successful in creating a cloud business.
First, the right executive leadership is in place in Bill Veghte, who took over as HP’s chief strategy officer following the retirement of Shane Robison last year. The important thing here is that Veghte wears another hat as Executive Vice President of HP’s Software Business. The combination of being a member of HP’s executive committee, owning direct responsibility for a major business group, and having oversight over corporate strategy is a powerful mix that seems (to me) to be right for the job. The fact that Veghte leads HP Software, home of powerful automation technology (think of the datacenter automation software bought with Opsware) resulting from years of patient acquisition, is another pointer. If any HP business needs cloud services to succeed, it’s probably Software, if only to provide a replacement income stream for the loss of on-premises deployments of longstanding products such as OpenView as customers move work to cloud services.
Another important point is that Veghte has been through the creation of a major cloud infrastructure before at Microsoft. In fact, I attended a number of Veghte-led sessions in the 2007-2008 timeframe when he briefed HP executives about Microsoft’s datacenter investments and the way that they planned to create the essential software base (using Windows) to host cloud applications. I imagine that this background and knowledge has been invaluable in charting HP’s path to the creation of its cloud offerings.
Second, HP has reasonable experience in building and running cloud services such as Snapfish and Magcloud. Issues such as how to scale and manage the storage necessary to accommodate the amount of data that cloud services typically have to cope with plus customer billing and delivery have been worked out since 2005. I am not saying that everything is perfect here as clearly HP’s cloud efforts to date have been spotty and inconsistent, but a track record does exist.
Third, HP Labs has been researching cloud services for a long time in an effort to understand aspects such as automation, security, and management. It’s a couple of years since I sat on HP Labs project review boards, but I imagine that progress has been made since and that HP Labs has made a contribution to what is now being offered.
Fourth, HP has an incredible amount of hardware assets that it can draw upon to equip its datacenters. Servers and storage will probably be highlighted here, but HP also has a very solid networking business based around ProCurve and some recent acquisitions such as 3Com and TippingPoint. Cloud businesses depend on solid networking and it’s nice for a cloud provider to have a complete networking business available. On the server and storage side, it’s also probable that HP has learned a lot from its provision of hardware to other cloud providers. The kind of servers used in cloud datacenters differ from standard ProLiants sold to the corporate world in that they are very much simpler and designed to be rip and replace compute boxes. The same is true of storage where massive racks of JBOD tend to be the preferred choice for cloud datacenters. I’m pretty sure that HP’s hardware teams have come up with some interesting technology for use in their datacenters.
Last, time was running out and it became imperative for HP to make a move into cloud services, if only to protect and enhance its reputation for the design and operation of large internal IT infrastructures that are now called “private cloud”. Remember, HP has to continue to sell a massive amount of hardware to corporations around the world to support its operations and this task would become a lot harder if HP couldn’t provide its cloud credentials. The same is true for HP Services, which became a massive managed services player following the acquisition of EDS in 2008. HP Services can now offer customers a range of public and private cloud services and this should help it better secure multi-year customer management contracts.
Of course there will be many bumps to navigate along the road to success. Amazon is already a big player in public cloud services and HP can expect to compete with solutions from IBM and other players. Microsoft, HP’s major partner in many other areas, will compete with Azure and Office 365. And some of the customer technical staff that HP has to convince might consider its chosen architecture (OpenStack and KVM) to be the wrong choices. Time will tell.
The net-net of HP’s plunge in public cloud services is that it’s a major strategic move for a massive company that they have been considering for over five years. We will soon find out whether the time spent waiting was a good move because it allowed the market to mature and all the necessary pieces to come together within HP or if the delay has compromised HP’s ability to succeed in what is becoming an increasingly competitive cloud services market.
Follow Tony’s ramblings on @12Knocksinna
So, finally HP made it into the public cloud hype! Bad for HP, because first of all they will face the same issues with public cloud custormers outside the USA as all others US public cloud provider do: Patriot Act. Need to say more?
Second, no company with a serious idea of how to operate their IT will give control over their data and infrastructure out of their hands.
Third, between the customer and his data in the public cloud ther is one little thing called The Internet, driven by Network Providers and their infrastructure equipment, especially DNS Servers. In case of a outage of that DNS Servers, no one can reach ANY server, despite of google.com or the servers of the customers public cloud (we had some DNS outages in Germany over the last months). But these network providers are often not also the providers of the public cloud ….. so what can a customer do in this case? Only wait until things are alright again?
Don’t think that this was a good idea ……
Well, I agree that the small but important component called the Internet remains a key element in any successful cloud deployment, but I think that the situation around cloud is more positive now than it was a number of years ago.
First, companies like Amazon and Microsoft have proven the utility of public cloud compute and storage services.
Second, there’s been a lot of experience with cloud outages that provide excellent signposts to companies who are considering using these services. Your example of DNS problems is a good example. I suspect that similar problems that afflict the underlying Internet infrastructure will decrease over time simply because more reliance is placed on them. I’m actually more worried about things like the Azure bug on Feb 29 that brought that service crashing to a halt.
Last, the U.S. Patriot Act is an issue for any company who doesn’t want their data to be held in any place where the U.S. authorities might be able to access it. U.S.-based companies like Microsoft, Amazon, and HP have to deal with this issue. The tactic of deploying datacenters in multiple countries doesn’t seem to help if the U.S. authorities really want to get to the data. However, I am sure that any company that is considering using cloud services will factor this along with other items into their decision process. Some will find it acceptable, others will seek a cloud provider that’s based in their home country. Others will simply take up the lessons learned in cloud deployments (mainly automation, simplification, and scale) and use them to create their private cloud infrastructures. I actually think that HP will benefit here because they can leverage their experience gained from the public cloud and use it with customers.
We definitely live in interesting times…
First of all I totally agree regarding to the PRIVATE cloud issue. Private Clouds with the possibilities of virtualization and provisioning are great! Second, Cloud Provider and Network Provider. The reason why the Swiss Telecommications Company Swisscom is relative successful with their public cloud offering is not only the fact, that they are strictly on Swiss law, but also that their customers can have booth services from them: the cloud and the network. Same is true in Germany for T-Systems.