Enterprise grade public cloud: IDC’s take 19 Jun 2013Posted by Tony Law in Cloud, Consumerization, IT marketplace, ITasITis, Managing IT, Tech Watch, Technorati.
add a comment
I’m on an AT&T webcast relating to public cloud infrastructure and its growth. Allow that this is primarly a US-focussed perspective. It’s AT&T sponsored, but delivered by IDC. It’s being recorded, and I’ll add the URL when it’s available.
Much of the underlying data comes from IDC’s winter 2012 CloudTrack Survey, with around 500 respondents. Five elements: the pace of change; deployment; networking; workloads; and next-generation solutions.
IDC refer to the “third platform”, not just second platform; and with spend growing nearly 12% per year compared to less than 1% for second platform. Third platform will account for almost 25% of this combined spend by 2020, and in the next three years spend on external services will grow to around an eighth of “traditional” IT spend. Over three quarters of North American companies are already using public cloud services.
There’s a useful categorisation of cloud deployment models, with names that speak for themselves. Self-run private or managed private; dedicated (externally) hosted or virtual private cloud; or public. Running across these are the decisions about on- or off-site, and dedicated or shared infrastructure. That eighth of spend shift over the next three years depends on these decisions.
Virtual-private cloud (VPC) has clout, through additional security and control, better connectivity into corporate networks, and more controlled SLAs but are built on public cloud infrastructure. AT&T believe shared services will command the lion’s share of the developing spend, although the split between dedicated and shared is more equal right now. This is what AT&T imply by “enterprise grade public cloud”.
Connectivity is crucial (remember, AT&T is a network company …) and there is an opportunity to connect VPC through an MPLS (multi-protocol label switching) high-availability cloud network rather than the public internet. Integration to the corporate network is close to seamless. IDC believe this option overcomes many enterprise objections to VPC cloud usage. And the CloudTrack survey suggests that any major workload coming up for reinvestment is at least going to be considered for cloud migration.
Noticeably, the workloads most likely to be moved are about the key elements of the “third platform”: social, big data (and analytics) and mobile. Where relevant, emerging markets also make a strong contribution to the importance of the third platform. Enterprises will need competencies across cloud and all these; they may not be tagged as cloud initiatives, but in these spaces cloud is crucial for developments to be effective, and those developments will be combinations of the four technology spaces. There’s a graphic for this; look in the webcast when it’s online (I’ll add the URL when it’s available).
On the half hour. Transition from the IDC analyst (Frank Gens, Senior Vice President and Chief Analyst) to Amy Machi, AT&T representative. This is a sales pitch for the combination of IBM’s Smart Cloud solution and AT&T’s VPN (NetBond), and you’ll get less notes. But with so much discussion about the limitations of service agreements with providers, it’s interesting that IBM trail over 70 auditable automated tasks available to clients, and cloud-based ITIL processes. Also, an important point is that AT&T will scale network capability in line with the demands on the scaleable cloud resource being claimed at IBM’s end of the wire. For anyone looking seriously at this version of the Cloud option, several case studies show the variation in possibilities.
Note, too, that at the present this is a US service and users need to be an AT&T customer. It will extend to Europe and Asia/Pacific relatively soon.
So: in response to questions, Frank Gens believes that investment in new capabilities will swamp legacy migration onto the third platform. And IT managers (VP/SVP) are coming to accept a reputable cloud service provider as having security at least as good as their own and possibly better, but the network has remained a vulnerability. With a managed MPLS network, rather than public infrastructure, these concerns are mitigating.
UK honours Apple’s designer 31 Dec 2011Posted by Tony Law in Cloud, Impact of IT, ITasITis, Technorati.
add a comment
Don’t suppose it’ll make most of the headlines, but Jonathan Ive has been awarded a knighthood (KBE) in the UK’s New Year Honours list. Though my paper, the Guardian, in their online report, only refers to him as “designer of the iPhone, iPad and iPod”. Well, we know better!
For interest, the Guardian offers a fully accessible list of the honorands. I found a couple of CBEs for computer science professors (one at what’s now Queen Mary, University of London, where I started my IT career), and a couple of retired Civil Service IT directors, but otherwise remarkably little recognising the UK’s information science capability.
Interesting that, to publish this information, the Guardian has created a Google Docs spreadsheet. No point in inventing your own infrastructure when the cloud can do the job!
• New Year honours list reflects my aims for ‘big society’, says David Cameron, Guardian, 31 Dec 2011
• New Year’s Honours, 2012 …, Guardian datablog, 31 Dec 2011
• 2012 New Year Honours, Google Docs
Beyond gmail: Google apps event with BCS 11 Oct 2011Posted by Tony Law in Cloud, Consumerization, IT is business, IT marketplace, ITasITis, Managing IT, Tech Watch, Technorati.
add a comment
I’m at a BCS North London event at Google’s London office, listening to presenters from the AppsBroker consultancy extend my understanding of how Google Apps work. We’ve passed through the background stuff about using cloud apps in general and now getting to the meat. If you’ve wondered, like me, what Google APIs can really do, then this is an as-it-goes posting; watch the space! Any errors in understanding or interpretation are mine, of course.
How to write a Google-extended app …
1 – Appscript; 2 – Gadget APIs; s – Data APIs
Just seeing the down side of everything being online rather than on the device; the demo’s gone down through being unconnected. Notwithstanding that I’m doing this on Google’s guest network,, the demo doc is, it appears, “offline”. Embarrassing, even when the demo’s working on a ChromeBook, which admittedly does reboot nice and quickly!
When it’s come back, we get a quick view of the script code inserted into a Spreadsheet to quickly create a form with follow-on technology such as mail-outs based on the respondent’s input, or sending update notifications when an online document is changed.
2: Data APIs, based on REST rather than SOAP (HTML based, IIRC, but can use other languages eg. Java/script .NET, …). Can for example use Data APIs to push data into a shared spreadsheet in real time from multiple users/locations/sources, but maintaining one version of truth.
Google App Engine and Cloud Storage will have a >99.9% SLA from November. Cloud SQL (see Google Blog last week) is under beta.
— adding to the interest level, we just had a fire evacuation and a quick tour of Eccleston Square with the fire marshals. Now trickling back – at least, most of us. I think some people have decided to duck out.
In the pipeline: Google Big Query: online dataset analysis – data mining/BI application. And something called the Google Periodic Table (there’s an extra column in the Transition Metal section …) which visualises the family of applications and extensions. Prediction, for example, can look at web traffic and draw interesting conclusions. Lots of searches on “sore throat” might signal the start of a flu epidemic.
Abbreviated in response to the disruption: Dalim, chair of the Branch, talking about governance. What changes with the cloud? Some of the controls e.g. for change management; assurance from third parties, and provider management; identity and access management (d0 you still have super users?) and monitoring; evolving technology, complexity and challenges. Dalim offers an app assurance checklist [see BCS NLB website in due course].
Q&A … references to Google’s global infrastructure capability; e.g. guaranteeing at least four copies of data on different continents (that is, replication like Lotus Notes used to do). Regarding data protection issues – Google can’t at present commit to (for example) segregating data into the EU though this is being worked on. The offering currently may not be appropriate for heavily regulated in-country enterprises e.g. some areas of government, finance. Google, though, takes the approach that they are not data owners; they are data holders, and would pass access requests to the data owners. And there are data online about which countries request legal discovery, how often, and when. From the security point of view, just a glimpse of the multiple levels of protection applied to data.
Thinking about a portfolio of services: Google Apps will integrate both on-premise (e.g. with AD) and other cloud services (e.g. a strategic partnership with salesforce.com). And there’s a commitment to back data out if a service relationship is terminated. Cloud, to Google, is short term contractable (e.g. 12 month; or a little as 1 month) – no lock-in.
• Google Apps (follow the links)
• Google App Engine, Cloud Storage and Prediction API are open for business, Official Google Blog, 11 Oct 2011
• BCS North London Branch: Past Events 2011 (you may have to scroll for this event; presentations are not yet posted but are expected)
• AppsBroker consultancy
“Cloud” has become a FUD word 7 Sep 2011Posted by Tony Law in Cloud, IT is business, ITasITis, Tech Watch, Technorati.
add a comment
A LinkedIn post flagged me to a Forbes report about a spat between Mark Benioff (that’s salesforce.com to you and me) and Larry Ellison (Oracle). About the definition, or the understanding, of Cloud.
Well, the first interesting thing about the report is that it’s not in some tech geek publication. It’s in Forbes, which rich people read. If ever there was a candidate for airline management’s key publication, it could be this one. It does rather confirm, doesn’t it, that Cloud (we used to say Cloud Computing) is mainstream business news.
And the second thing is that it confirms, as we already knew, that Cloud has become one of those Humpty Dumpty words. You know: When I use a word (said Humpty Dumpty to Lewis Carroll’s Alice) it means exactly what I tell it to mean, neither more nor less. It’s happened in every IT generation. Working backwards, we certainly include Grid, we include “e” (as a prefix, such as “eServerFarms”), and we probably include client-server. And more, I’m sure.
As an adviser, facilitator and consultant I need to understand what people are thinking when they say “Cloud”, and it can be a lot of things these days. It’s my perception (and I’m by no means alone) that a lot of what’s marketed as Cloud today is one of:
• old-fashioned hardware-based outsourcing to a remote data centre
• web services
• some newer form of outsourcing
always with long term contracts, fixed prices, security, and and and …
We can do better. But first, there are a couple of things Cloud doesn’t need to be.
It doesn’t have to be “cheap”. This is a benefit in many cases, but not a fundamental. And in any case it’s relative: a service used for a short period may be expensive per unit, but still cheaper overall than provisioning your own “stuff” which you have to lay in for the long term. A comparison: taxi fares aren’t “cheap”, but if you don’t need permanent access to your own car then occasional taxis have the edge over the long term capital and recurrent costs of running one. But the key point is: no payment in advance, no commitment to spend levels, no true-up.
And it needn’t be “public”. I’m perfectly happy to include what are called “private cloud” services in the definition, so long as they are still true Cloud by the criteria below. But the key point here is: Cloud is not just a new word for a conventionally provisioned in-house data centre.
Many, many service vendors are rebranding their outsourced or managed services as “Cloud” to cash in on the hype. There’s a massive overlap between what we consider “virtualised” and what we consider “Cloud”. And service buyers are adding to this by insisting that cloud services must be as secure, stable and long-term an investment as any other outsourcing deal. Fear, Uncertainty and Doubt ride again.
Some (many) years ago, I was part of the team operating a then-new ICL 2980 for London University. The “V” in “VME/B” stood for “Virtual” and we had to learn (and explain to the users) the differences of a virtualised system and the advantages it could offer in the way they approached its use. Yes, this was the totally modern 1980s. Other operating systems were “going virtual” too, and one of the trade papers (I think it was Computer Weekly) ran a definition I’ve always remembered:
If it’s there, and you can see it: it’s REAL
If it’s there, and you CAN’T see it: it’s TRANSPARENT
If it’s NOT there, and you CAN see it: it’s VIRTUAL
If it’s NOT there, and you CAN’T see it: it’s GONE.
I think we add one more:
If it’s NOT there until you WANT it: it’s CLOUD.
And here are my criteria for a service to be called Cloud:
• accessed over the network using Internet protocols
• available immediately on demand
• de-provisioned immediately after use
• easy sign-up
• no long term commitment to the service provider …
• … nor by the provider to the customer
• payment strictly by usage metering
• payment after the fact, not in advance
• as near infinitely flexible capacity as can be
• Larry Ellison and Marc Benioff Just Can’t Agree: What Is the Cloud? Forbes, 6 Sep 2011
• ICL VME, Wikipedia
Total Economic Impact: a full case study 12 Jul 2011Posted by Tony Law in Cloud, Insight services, IT marketplace, ITasITis, Managing IT, Technorati.
add a comment
Forrester’s Total Economic Impact methodology has been around for some time now; Chip Gliedman and his colleagues must have trained many IT executives in its application, its models, and its succinct elevator business case (“We are doing X to make Y better, as measured by Z, which is worth $N to the company”).
But training, and the examples within the training course, only go so far. A note in one of my regular alerts led me to a full case study, commissioned from Forrester Consulting by Cisco. It’s a valuable read. It’s a fictitious case, of course (“Company A” style), but it goes right through the whole analysis, as well as including a summary of the methodology itself
Cisco wanted to illustrate the case for their “Borderless Networks” technology. I’ve long been a believer in this methodology, not least because it captures opportunity costs and benefits, and puts cash value on risk. So it’s interesting that Cisco chose TEI as the vehicle to make this case.
Read it if:
• you want to know more about TEI than is available in Forrester’s summaries; you have to pay for the training course (or get access through a subscription)
• you know about TEI and want to see how a full evaluation works
• you’re looking at facilitating what Cisco describe as “secure, reliable, and seamless connectivity to any device, [to] any person, and in any location”
You’d need to research the Borderless Networks concept separately; this document is about the business case, and might well be useful independent of Cisco’s own technology.
• Total Economic Impact of Cisco Borderless Networks, Forrester Consulting, November 2010, available through Cisco website (no registration required)
• The Total Economic Impact Methodology: A Foundation For Sound Technology Investments, Chip Gliedman, Forrester Research, 4 Aug 2008 (subscriber access or purchase)
• Borderless Networks, Cisco overview