Categories
Alan Weissberger Cloud Computing

Cloud Leadership Forum: Opportunities & Obstacles to Cloud Adoption

The just concluded Cloud Leadership Forum (June 20-21 in Santa Clara, CA) brought together CIOs, IT managers, vendors, and analysts to discuss the current status, future directions and important caveats of Cloud Computing. This very successful IDG event was organized and chaired by IDC Chief Analyst Frank Gens and IDG Enterprise Sr VP John Gallant. Both did an excellent job of running the conference, which provided very useful information to the many attendees who came from all over the world.

A IDC/IDG survey revealed that information technology executives believe that cloud computing will have significant impact on IT organizations and IT vendors, as well as the enterprises they support.

Frank Gens of IDC is a strong advocate of Public Clouds, insisting they will dominate over Private Clouds sooner than most expect. He says that IT executives need to be looking at cloud for the new capabilities it offers, not simply to re-create what the business is already doing on a potentially cheaper platform. The implication is that most new apps – even those that are mission critical-will run on data centers in the cloud.

All but the largest companies appear to favor Public Clouds because of lower cost, advantages of scale, and faster depoyment. Amazon, Rackspace, Microsoft and now Fujitsu all offer public clouds for Infrastructure as a Service (IaaS). However, HP was touting Hybrid Cloud (mixture of both public and private) as the way forward. HP plans to introduce their own public cloud later this year.

Here are just a few data points from the Conference:

  • More than 70 percent of those surveyed said they believed that by 2014, a third of all IT organizations will be providers of cloud services to customers or business partners.
  • Almost 80 percent of respondents felt that cloud service brokers and service aggregators will provide integration, management, security and other services across public cloud offerings by 2015.
  • More than 80 percent of the respondents said that one-third of Fortune 1000 enterprises will deploy at least one business-critical system in the cloud. More than half of the IT executives surveyed believe that mobile-optimized cloud services will be a primary interface with customers by 2014.
  • In 2010, more than 40% of storage purchased was for cloud use (cloud computing and ISP internal use). This percentage is expected to increase this year and next, especially in light of Netflix’s move from in house to cloud resident data centers. An IDC analyst suggested that shared cache memory might be used in the future for cloud storage.
  • VCs are not investing in compute/storage/networking hardware companies or infrastructure players. It’s believed those mkts will be dominated by large players who have massive scale. Their investments are in cloud, e-commerce, social media start-ups that have differentiating IP.
  • Platform as a Service (PaaS) continues to be the most unexploited area of Cloud. This is because the larger players are all doing Infrastructure as a Service (e.g. Amazon, Rackspace, Fujitsu, IBM, HP, etc) while Sales Force.com and start ups dominate Software as a Service.
  • While the advantages of cloud services are very well known and have been hyped to the sky, several speakers emphasized business innovation and agility as key benefits. Their message: use the cloud to accomplish business objectives and innovate; adopt to changes quickly.
  • A key opportunity that Frank Gens of IDC has emphasized is the “mashup” or integration of cloud services with big data, business analytics, mobile device, mobile apps &data, and social media. Due to the variety of platforms and lack of standards, we think this will take a very long time to happen on a large scale.
  • Security, availability, ;ack of standards =vendor lock-in, guaranteed performance, contracts and SLA compliance seem to be the key impediments to massive cloud adoptions.
  • Cloud standards are nowhere on the horizon. What’s urgently needed is a common API to access cloud services (see discussion below). That and other critical cloud interoperability standards will likely take at least two more years to be sorted out.

What’s the Impact of No Cloud Standards?

No standards translates into a lack of interoperability, which locks-in the customer to a single cloud service provider. For example, the lack of a standard IaaS management API requires cloud user software to be particularized to a single Cloud Service Provider. This is because every cloud service provider has a different version of web services (mostly REST based, but a few use SOAP WS’s) protocols and software interfaces that the customer must adhere to.

Here’s an article highlighting IEEE CIOs position that lack of Cloud Interoperability is a greater problem than security for most potential Cloud users:
http://community.comsoc.org/blogs/ajwdct/ieee-cio-says-cloud-interoperability-bigger-problem-security

In the absence of any Cloud network (UNI or NNI) standards, there are a variety of network interfaces (L1-L3), which are highly dependent on the type of network used to access the cloud services.

  1. When the public Internet is used to access cloud services, there is usually an IP VPN per user with IP-SEC or IP-MPLS/PPP frames over whatever the physical medium dictates as the PHY layer. The same Router that is used to access the Internet is used to obtain cloud based services (usually public cloud).
  2. But if a private network is used to access the cloud service(s), the network interface could be IP-MPLS, MPLS over Ethernet, Ethernet VPLS, or other L2 protocols over the PHY layer. This means the private network provider (not the corporate end user) is responsible for the on premises access switch/ router that is used to accesses cloud services. In most cases, e.g. Savvis, AT&T, Verizon Business, the private network provider is the same company offering cloud services. But in any case, if the customer switches cloud providers they will need to install a different access switch/router. Note also, that switch/router may not provide access to the public Internet. It sure sounds like vendor lock- in to me!

Other Problem Areas for Cloud to Overcome

There are other pitfalls and barriers to cloud adoption besides the lack of standards. While everyone knows that security is an impediment to cloud adoption, other important functional (vs contract and legal) issues are reliability, availability, fast failure recovery (e.g. the recent Amazon EC2 failure took up to 5 days to restore service), performance (especially when the public Internet is used to delver Cloud Services).
Finally, there are some BESIDES this author who say that the network is not ready for cloud and will be the bottleneck:

” Networks are not prepared for cloud onslaught. Complexity, latency, bandwidth. Networks just are not prepared, particularly as organizations build public-like cloud services in their private clouds. Virtualization at scale or greater numbers of devices on the network don’t help either. Time for a cloud-ready data center (Juniper example, but there are many). One challenge is that cloud is reducing capital expense budgets, where network upgrades happen. Are we too early for that?” –Paul Calento 06/22/2011 – 3:09pm.

“I fully agree. Yes Paul is right! THE network (aka cloud) is the critical infrastructure, but you can indeed get the connections you need and prices are falling however it strongly depends what application. A good example is VDI this was sold as the hottest thing the last 2 years, but you will not be able to run this in the cloud! Once we all have fiber to the desk maybe!” Oliver Fischer-Samano on Wed, 06/22/2011 – 3:43pm.

Here’s a related article about John Gallant – the co-host for Cloud Leadership Forum the last two years:

Cloud Computing Conference Leadership from a Master IT Journalist, Content and Editorial Director

Note: This author is closely tracking cloud standards being produced by the ITU-FG Cloud as well as the IEEE P2302 Cloud Interoperability Working Group.

Author Alan Weissberger

By Alan Weissberger

Alan Weissberger is a renowned researcher in the telecommunications field. Having consulted for telcos, equipment manufacturers, semiconductor companies, large end users, venture capitalists and market research firms, we are fortunate to have his critical eye examining new technologies.

8 replies on “Cloud Leadership Forum: Opportunities & Obstacles to Cloud Adoption”

Thanks for reporting on the conference, Alan and thanks for keeping us up to date on what is going on with the cloud. It seems like the cloud will become increasingly intertwined in telecom providers’ offerings. Look no further than last week’s purchase of Zerigo as an example of an entity that wants to add value to its offering for its existing customers. I am sure we will see more and more of this type of activity.

Ken, Thanks for your comment. Agree with your observation that Cloud will become part of telco offerings, i.e. the telco as cloud service provider. However, only the telcos that have experience with IT Hosting (compute servers and storage) will likely succeed. In the U.S., that short list includes Century Link (which has acquired Qwest and is acquiring Savvis), AT&T, and Verizon Business (whose cloud service is being provided by its Terremark subsidiary). According to Light Reading: “Core to the Verizon strategy will be helping customers assess what applications gain speed to market or operating efficiencies by moving to the cloud, and which ones continue to sit in the traditional IT space or in collocation sites in between, and then providing services that connect, secure and scale this broad range of IT services and apps.”
http://www.lightreading.com/document.asp?doc_id=209322&f_src=lightreading_gnews

Sprint is not in the cloud picture and is focusing on Machine to Machine (M2M) communications using its 3G/4G wireless network. Smaller telcos only have experience in back end management systems (OSS/BSS) which will not be particularly helpful in evolving to cloud based services for business customers.

Another huge issue wireless telcos face is mobile data roaming and network availability/ coverage. This is a general concern, but specifically with cloud access by a mobile workforce. As an example, how does an executive with a 3G iPAD from AT&T, access cloud services when he is in VZW or Sprint territory? And what speed, throughput and latency will he get?

Here is what Rethink Research writes about Mobile Cloud:

“Mobile cloud services are emerging that synchronise data across multiple mobile devices with centralised storage in the cloud.

ARCchart forecasts that demand for mobile cloud storage will grow exponentially as users increasingly adopt these services and the volume of mobile content they generate cumulatively grows. Photos and video captured on mobile devices will overwhelmingly carry the greatest cloud storage demand and the average size of these media files will grow substantially as camera pixel resolution continues to increase.

We predict that mobile-generated content will consume 9,400 PB (petabytes) of cloud storage by 2015.”

Opinion: This can only happen in the U.S. if there is roaming between 3G-HSPA, 3G-EVDO, 4G-WiMAX, and 4G-LTE. And later, true 4G= LTE Advanced.

One of the impediments to mobile cloud and mobile computing has been high data roaming charges. Several newspapers today report that Europe will drastically reduce those charges, which will make true mobile computing more attractive for users. Note, however, that in the U.S., high data roaming charges or unavailability amongst different 3G/4G networks is still a barrier for mobile computing and the mobile cloud.

The WSJ reports: “Mobile operators in the European Union will be ordered to cap the costs charged to customers using their phones to access data services while abroad, EU digital commissioner Neelie Kroes is expected to announce Wednesday.

Such a proposal would come after the European Commission last week proposed to extend roaming caps on calls. The proposal must also be reviewed by the European Parliament and the EU Council, which brings together heads of government for the 27 EU countries.

In a draft legislative document, the commission says, “The roaming market is not yet competitive and further regulatory intervention is required.”

Currently, data roaming is subject to a maximum safeguard cap of €50 ($73), but starting July 1 next year it would be capped at 90 European cents ($1.31) per megabyte, according to the draft. The cost ceiling for data used would decrease to 70 cents a megabyte in 2013 and 50 cents a megabyte 2014, and would remain at that level until June 30, 2016, when the regulation expires.”

http://online.wsj.com/article/SB10001424052702304760604576425882975782852.html

“In 2010, more than 40% of storage purchased was for cloud use”

wow i should really look into this, it seems more expensive than conentional hosting, even though it is scalable

Is it possible prices will reduce sooN?

At the Connections Digital Living conference in Santa Clara, CA today, it was hinted that consumers would have personal clouds for storage, music, videos, etc.

Later, Real Networks stated they just announced a new product that maintains all the music/videos/photos from every device in a subscribers home on their Internet site. That’s sort of a personal consumer cloud, isn’t it?

Any comments regarding privacy and security of such a personal consumer cloud?

The Real announcement at the Connections Digital Living conference was significant, as it is about controlling one’s data and putting the consumer in the middle of that data.

One thing that I wasn’t able to confirm, but looks very interesting, is that it looks like Real may have figured out a way to capture the meta-data and data (e.g. photos) that one puts on Facebook and other social networking sites, so that it isn’t lost if a given social media network dies or changes their terms of service.

It will be interesting to see how widely adopted the Real product, Unifi, becomes.

Affinegy also had a similar concept that they showed at the conference that allowed one to create a “personal cloud”. This is a service provider-centric offering.

http://www.viodi.tv/2011/07/06/leading-the-orchestra/

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.