Cloud Performance-monitoring - 0 views
started by Engel Tange on 11 Dec 13
started by Engel Tange on 11 Dec 13
Start a New Topic » « Back to the Steps to make a Pot Dabber group
started by Engel Tange on 11 Dec 13
There is much evidence to back up such analogies, if current trends hold, as they certainly look like. Consider:
Every customer with a smart-phone, every Web-user with a Gmail or Hotmail address, every Facebook friend is earnestly involved in the cloud.
The U.S. government is moving hundreds of its data centers for the cloud. The Pentagon has numerous cloud projects underway that will eventually touch every soldier in the field, and has already equipped Special Operations Forces with Android-powered tablets.1 NATO is piloting an IBM-built program at its Norfolk, VA facility, and hopes to roll it out to any or all 28 alliance nations.2
In the world of business, adoption of software-as a service jumped from 66 percent of businesses in 2007 to 90 percent in 2010. 3
To the supplier side, major companies are snapping up smaller data storage companies with fervor similar to the dot-com boom. Verizon this cold weather announced a $1.4 million deal to obtain Terremark World wide. And Hewlett-Packard, EMC, Dell, and IBM have all ponied as much as buy 'a multitude' of data-storage businesses, according to a New York Times report.4
The tech giants are all investing huge sums in their own data centers. Microsoft CEO Steve Ballmer announced last fall that the company will spend billions in cloud data centers.5 Apple is opening its billion-dollar data center in New York this spring. IBM is planning development of the largest data center in China - nearly 6.7 million square feet.6
And in a veritable land rush in the Pacific North-west, Google, Microsoft, Yahoo, and now Facebook have either developed or plan to develop data centers, incentivized by local tax-breaks but more importantly, to become close to abundant supplies of cheap electricity, much of it supplied by the Columbia River.7 Data centers are such power hogs that Google formed an unique power business, Google Energy, to enable it to get and sell power in the regulated wholesale market.8
Last but not least, there's the cash - plenty of it. On the consumption side, businesses are seeking to enjoy major IT savings by moving information and processes towards the cloud. To the delivery area, there's plenty of money to be made in satisfying that want. Amazon Web Services is expected to possess profits exceeding $2 billion in 2014, and is likely to grow 50-percent in 2011, to $750 million. RackSpace Hosting, another key player, had 2010 revenues of $100 million.9
In Europe, the Centre for Economics and Business Research projects that cloud processing will produce more than €763 billion - that's more than US$1 billion - in 'final financial gain' (which includes IT savings) from 2010 to 2015, and directly or indirectly create more than 2.3 million net new jobs.10
A ready and fairly sensitive industry, quickly developing infrastructure, amazing financials - the level is set for that cloud to ultimately change the way the planet shops, manipulates, and accesses data.
Simple Strategy, Complex Meanings
Confident Google CEO Eric Schmidt is credited with explaining the vision, in the past in 1993, that became a mantra for Sun Microsystems, where he was then CTO: 'The network is the computer.'11 In its most elementary type, this is the substance of the cloud - that the collective resources of processing power, storage, application and network are able to be harnessed anytime, everywhere, by users to whom it can not matter where anything is situated.
The National Institute of Standards and Technology is here at this more complicated official description of cloud computing: 'a model for allowing convenient, on-demand network usage of a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) which can be quickly provisioned and introduced with little management energy or company interaction.'12
'It is a definition merely a standards human body might love,' states Anshu Agarwal, vice-president of marketing for Keynote Systems. 'The way we consider it instead is that 'the Web is now the cloud.'With the open Web, open Internet standards and methods, and the common denominator of http, it's advancing interconnectedness at a pace that's truly breathtaking.'
In layman's terms, for customers and business system end-users, the cloud means your electronic stuff doesn't live inside the box (or smart-phone or supplement) that's in front of you. It is someplace else. For enterprises, it means some or every one of the corporate computing power and data has moved out of the corporate data middle to somewhere else - and 'somewhere' else can mean a number of sites (see sidebar).
The Four Kinds of Cloud
In meteorology, you have cumulus, cirrus, stratus, nimbus. In processing, you have private, public, hybrid, and community. Here's what each means as defined by NIST.
Individual cloud. The cloud structure is run exclusively for a corporation. It may be handled by the business or a third-party and may occur on-premise or off-premise.
Public cloud. The cloud structure is created available to the general public or a large industry group and is owned by a corporation selling cloud services.
Hybrid cloud. The cloud infrastructure is a arrangement of two or more clouds (personal, community, or public) that remain unique entities but are bound together by standardized or proprietary technology that allows information and application portability (e.g., cloud exploding for load-balancing between clouds).
Area cloud. The cloud infrastructure is shared by many organizations and supports a certain area that has shared problems (e.g., mission, security demands, policy, and compliance concerns). It may be managed by the organizations or a third party and may occur on premise or off premise.
Source: National Institute of Standards & Technology, http://www.nist.gov/itl/cloud/upload/cloud-def-v15.pdf
The Cloud's Silver Li-ning
How come reaching for the clouds this type of hot topic in the business world? Underneath it all is that this irksome statistic: Some 70 to 80 % of IT budgets are dedicated only to maintaining the lights on - maintaining electronics, pc software and network connections, and offering people. The lion's share of IT budgets is sucked up by the status quo, perhaps not for adding functionality, streamlining operations, or gaining a competitive edge.
The cloud promises to improve that and provide extraordinary benefits in cost-savings, elasticity, agility and simplicity.
Financial savings. The price savings from cloud computing will come from several recommendations. The cloud puts the cover on CAPEX: The expense of maintaining and buying computers might be eliminated (which is why many organizations think about the move when their current hardware is near end-of-life). Pc software preservation - repairs, patches, changes and improvements - is held by the cloud service, reducing IT staff time and cost. And those important energy costs mentioned above get to be the cloud provider's price also.
Agility. Freed from a lot of the maintenance problem, IT staffs could focus more on developing capabilities that answer market changes, increase efficiency to functions, and develop competitive advantages. And they can appropriate cloud resources due to their development work and screening, without provisioning in-house servers and bandwidth. The result is faster 'time to value.' Furthermore, if the organization requires more computing horse-power, it is an easy, rapid, self-service process via a Web portal to add it.
Strength. Retailers need remarkable capacity in November and December that will lie idle the others of the season. Accountants hit their peak in April and March, together with the other ten months quiet in contrast. 'Traffic typically picks up in the end of Q3 and in Q4, typically 4-6 times higher than an ordinary week or month,' says Alex Taranukha, senior director of architecture for the website ShopLocal.com. 'Instead of putting in more boxes, we just add more service to your load stability from Rackspace.'
And for CPU-intensive business processes, strength often means having an army of machines open to quickly complete tasks that would take much longer on a company's own, smaller, inner variety of machines.
Simplicity. In case you require to learn more on website monitoring, there are heaps of resources you could pursue. Nothing goes out-of-date quicker than computer hardware and software. Shifting the frustration of keeping everything up-to-date to the cloud supplier is a very attractive proposition. Nonetheless it is not without trade-offs. 'Under certain circumstances, the cloud-based services are often less feature-complete compared to conventional [boxed] software,' says Todd Kelley, VP of operations and managed support services for Competitive Computing (C2). 'It is less configurable or custom-made, and therefore some businesses are experiencing that trade-off - it costs less and is more scalable, nevertheless they can't always do anything their own special way.'
It's Not All-blue Skies And White Clouds
That is a fantastic list of possible benefits, but enterprises in general and CIOs in particular also provide serious concerns about trusting their digital functions and data to third parties largely beyond their control. Protection and data privacy is chief among them. Many businesses have invested heavily in security structure and methods, and are understandably dubious of a cloud provider's ability (or willingness) to match their efforts. Co-mingling of information with that of other programs in co-tenancy scenarios is likewise a concern.
Reminders of weaknesses within the cloud appear regularly inside the daily headlines. Recent case in point: Google loses most of the data for 150,000 Gmail account consumers in the course of performing a storage software update. What CIO does not observe that headline and picture his or her own users' a reaction to such a problem? And how can it get solved if it is all out of IT's reach and get a handle on?
Regulatory compliance is yet another area of concern for those in certain sensitive sectors and publicly-traded organizations. Speaking about private-practice medical expert communities, C2's Kelley tells Benchmark, 'We've options where they could have all [their records] off-premises within the cloud - options that cost less - but they don't want a lot of anything regarding it.'
Also accepting safety and compliance issues might be satisfied, there is still an enormous elephant in the area that threatens the achievement of cloud-based operations: performance.
Performance: The Cloud's Achilles Heel?
Think of all of the methods in a hypothetical electronic value chain. An order is available in via a branch office or store. Headquarters is queried for the customer account data. The manufacturer (third-party) is queried on stock and delivery schedules. A request is sent to the logistics company to take delivery from the company and deliver to the consumer. Each of the data extends back to corporate to get ready the account and update the client records. Then the paperwork is repaid to the originating branch office or store.
In the days of the past, any of the orders happening outside the branch office or corporate data center would require emails and telephone calls. Today, though, everything travels over the Web - the cloud - via both direct connections and Web APIs. The headquarters data center, and those of the manufacturer and logistics company, may or may not even be on their premises. They may be a virtual installation to the banks of the Columbia River.
How will you evaluate and monitor performance under such a situation? If there is an issue, how would you pinpoint it along this convoluted chain?
Field agents could be used to measure performance to the person anytime, anywhere, as Keynote has been doing for a decade and a half, if it had been simply a website. If everything happened inside the corporate information center and within the company's private WAN, dimension agents might be installed at appropriate locations. But today's digital value chain is both of these and more.
'In the old company world, you'd mount either hardware or software agents, 'hearing posts,' throughout your structure, throughout your infrastructure,' Keynote's Agarwal says. 'But that you don't own the whole environment today. And your partner isn't allowing you to place small listening agencies in their infrastructure, so you really do not have visibility anymore. The only time you understand something goes wrong is when programs are a deep failing for consumers. And how can you possibly find out what is gone wrong and where if you don't have visibility to the whole transaction sequence'?
Whether it is a personal, public, or hybrid cloud, IT still is held accountable for performance. And so IT has to get ways to observe all of these links in their digital value chain is doing, even though it involves the black boxes of cloud suppliers and/or inaccessible partner/supplier information stores. That means having providers tracking every link in the chain - at point of firing, inside and outside the corporate firewall, and at the very first and last mile. It's more complicated, but achievable - and imperative.
'Our public structure, our world wide monitoring network is quite well-positioned to monitor programs whether they are hosted in a public cloud or in a personal cloud,' says Keynote's Ian Withrow, senior product manager. 'Now we are able to get in the firewall and check each of the point-to-point links and traffic within a network.'
The most popular retail deal-aggregator ShopLocal.com has this sort of complex architecture and has put in place an equally sophisticated tracking system started by Keynote.
'Using Keynote's lightweight individual application agencies helped us too much to monitor first-mile performance and API endpoints. And at the same time, Keynote's public monitoring system let us observe Akamai and the last-mile efficiency the end-user sees,' says ShopLocal's Taranukha. 'So it actually gave us very great insight in to how different parts of the architecture accomplish, and helped us to allocate our hardware and resources at Rackspace, our hosting company, very efficiently.'
The Performance Cost
Minute by minute, degraded performance exacts a painful cost when mission-critical operations may take place. Study of IT professionals conducted by consulting company Enterprise Management Associates indicates the price of downtime 'clusters' around $60,000 an hour or so. Some 40 % of those interviewed reported 30 minutes or more of unscheduled downtime in just about any given month. 13
'For firms with critical programs, our typical estimate is $100,000 a minute,' says EMA Vice President Dennis Drogseth, incorporating, 'I was talking to your group about downtime for financial transactions. It had been one million dollars a minute. Which means you probably would not fully rely on an external cloud supplier for that app'
SLAs as Elusive as a Cloud
One reason cloud performance monitoring is so essential is the fact that cloud supplier service is so unclear. 'Performance' to cloud providers typically means only availability; and even availability is only loosely assured. For Amazon Web Services, being an instance, unavailability means no connectivity at all throughout a five-minute period; in case your user includes a bad, erratic, miserably slow connection, so far as Amazon is worried, they've delivered. And availability means availability when it leaves Amazon's door; whether or not anything actually reaches your user isn't Amazon's problem (no matter their options for ISP and connection). Oh, and the duty of proof is for you, the consumer. For all intents and purposes, if you even have company Amazon is not even checking to see.
This is not to dump solely on Amazon. The same guarantees, or lack thereof, are typical of many cloud providers. As well as the caveats above, scheduled and emergency downtime is excluded from access guarantees; charges for unavailability are small, and certainly not commensurate with prospective business damages; and any other sort of performance isn't contained in the service level agreement.
'Many cloud service providers don't offer such a thing beyond junk for SLAs,' Drogseth says, (see accompanying interview). But, he continues, 'Service level is critical. If you're a company and your cloud provider isn't investing service levels that you feel confident with, kick them out. The industry may change to a thing pair of companies who don't care about you. A more partner-oriented group that will begin to give you visibility as you require it - that's the group you want to use.'
A perfect cloud-client working relationship includes large SLAs, external monitoring of SLA variables that's visible to both supplier and consumer, and substantial recourse if the service falls short. Instead of this ideal, but, the burden is on the enterprise to put measurement and monitoring in position and to maintain their provider accountable - so they can either obtain the service level they require, or switch to an improved provider.
End-users do not fundamentally care where or how their purposes are delivered; they just care that they work. And individuals they hold responsible would be the IT staff.
'In the end,' Drogseth claims, 'your products and services or programs as an IT business - whether you pick them through services or whether you produce them, whether you deliver them partly through the cloud or not -will stand or fall on the success of the quality of your application services and your capability to deliver them cost-effectively and responsively.'
Long-Term Prediction
Given the 'cloud' moniker, meteorological analogies are correct (if overused, even here), and the predictions are similarly most useful confirmed by a look out the window. But given the tremendous investment happening in cloud infrastructure, and the case for offloading computing to the cloud, it's safe to say that cloud computing is a snowballing phenomenon that's planning to take over an important part of the world.
The system has become the computer. It's as much as business clients to determine, through rigorous performance-monitoring and measurement, whether it will be a super-fast, user-pleasing, productivity-boosting workhorse, or a mediocre-performing, user-annoying, productivity drag.