Hewlett-Packard on February 24^th opened an advanced collaborative research lab in Singapore to support its growing cloud business – what it calls its “Everything as a Service” vision.

“Our new lab in Singapore is a key component of the transformation strategy put in motion less than three years ago to accelerate our pace of technology transfer and bring a variety of advancements to market,” said Prith Banerjee, senior vice president, Research, and director, HP Labs, in a press release. “HP Labs Singapore aligns very closely with HP’s strategic growth areas and significantly expands the resources we bring to bear on our customers’ biggest opportunities and challenges.”

Officially known as HP Labs Singapore, the facility will focus on a range of projects that aim to re-examine data center and application design principles in order to explore how future cloud computing needs will be met, said the release.

HP plans for the lab to work closely with customers, partners, HP business divisions and schools to generate advancements that will drive research for cloud development. And HP expects its customers to capitalize on this shift to a service-based infrastructure model using developments from HP Labs Singapore.

The new lab will collaborate with other key cloud initiatives already underway at other HP Labs sites, including the Service Automation and Integration Lab (SAIL) in Palo Alto, CA., and the Automated Infrastructure Lab (AIL) in Bristol, UK.

HP’s intention is that, together, the three labs will work on its vision for creating an enterprise cloud software platform – dubbed Cirious. As part of that project, HP has partnered with:

– Intel, Yahoo! and the Infocomm Development Authority (IDA) of Singapore to create a global, multi-data center, open source test bed for the advancement of cloud computing research and education. IDA houses one of nine test bed locations worldwide.

– SingTel – to form Singapore’s largest commercial grid services platform. It’s called Alatum, and it offers a variety of computing power, storage and software applications on a pay-per-use, on-demand and online basis.

HP Labs Singapore is the company’s third research facility in the Asia/Pacific region and its seventh worldwide

Of course, HP isn’t the only one building data centers and research facilities around the world dedicated to cloud computing and development. It’s good to see that healthy competition in this area is further driving hunger and interest for cloud computing – and that, in turn, lifts all the cloud service provider boats in this giant industry.

With the increasing reliance on consolidated data centers, virtualization, and cloud computing, network management has become less concerned with the connections between individual network nodes and more concerned with overall application performance management. Network managers must manage not just network infrastructure, but also the performance of network applications.

CA Inc.’s acquisition of NetQoS Inc., a vendor of network and application performance tool, for $200 million clearly anticipates this trend.

Abner Germanow, IDC research director, says the CA aims to give enterprises a more holistic view of the network, something that traditionally hasn’t been handled by network management tools. A network manager’s toolset looks rather like a carpenter’s toolbox, Germanow says, with different tools for different users. NetQoS completes CA’s already respectable toolbox, he says, by adding more functionality such as the ability to manage WAN connections. This will make CA’s view of the network more application-centric, he says.

CA’s Wily family of application performance management tools has a solid reputation in the field. Unfortunately, CA’s products provide more data-oriented views of performance rather than application-oriented views. To best understand the performance of distributed and outsourced resources, such as cloud services, an organization requires both data-oriented and application-oriented views.

Federal agencies are excitedly ‘upgrading’ to cloud computing through today’s launch of Apps.gov. The new website brings the promise of cloud IT services to all federal agencies, and the hip new app service includes everything ranging from professional business applications, cloud IT services, productivity apps, to the ever so popular social media apps! With many of the services being offered by either Google or salesforce.com, the movement to cloud computing is expected to reduce government operational costs as well as increase productivity within Federal agencies.

Cloud computing, in which information and software are stored in centralized units that can be accessed by any number of devices simultaneously, is the next big thing in IT according to a blog authored by Kundra. He went on to state that the website Apps.gov is an all encompassing site for cloud services, due to the way it consolidates services which are currently available. This is set to revolutionize the operation of IT and save taxpayers money.

A whopping $75 billion is spent by the government each year on information technology, says Kundra at the Whitehouse bog post. Yet the procurement processes and security procedures prevalent in this time and age can use up a lot of time that could have been otherwise used in other endeavors, and are even redundant at times. Kundra writes that the practices of the past have often ended up in an inefficiency across the federal government of purchased IT capabilities. Kundra adds, ‘These barriers will be addressed - it is a priority. We will improve the methods employed by the government in leveraging new technology’.

Due to release in the year 2010, Google is working on a “government cloud”, also known as a “G Cloud“. The program will work in much the same way that Google Apps works, though will be made to suit the US federal state and local governments. FISMA (Federal Information Security Management Act - stating procedures on running federal government information applications) approval is currently being sought by Google for their Google Apps software. According to Matthew Glotzbach, director of product management for Google Enterprise, the process is close to done and soon we will know whether Google will be authorized to use FISMA.

Infrastructure as a Service, or IaaS, gives business access to vital web architecture, such as storage space, servers, and connections, without the business need of purchasing and managing this internet infrastructure themselves. Because of the economies of scale and specialization involved, this can be to the benefit of both the business providing the infrastructure and the one using it. In particular, IaaS allows an internet business a way to develop and grow on demand. Both PaaS and SaaS clouds are grounded in IaaS clouds, as the company providing the software as service is also providing the infrastructure to run the software. Choosing to use an IaaS cloud demands a willingness to put up with complexity, but with that complexity comes flexibility. Amazon EC2 and Rackspace Cloud are examples of IaaS.

Platform as a Service (PaaS) clouds are created, many times inside IaaS Clouds by specialists to render the scalability and deployment of any application trivial and to help make your expenses scalable and predictable. Some examples of a PaaS system include: Mosso, Google App Engine, and Force.com. The chief benefit of a service like this is that for as little as no money you can initiate your application with no stress more than basic development and maybe a little porting if you are dealing with an existing app. Furthermore, PaaS allows a lot of scalability by design because it is based on cloud computing as defined earlier in the article. If you want a lean operations staff, a PaaS can be very useful if your app will capitulate. The most important negative of using a PaaS Cloud provider is that these services may implement some restrictions or trade-offs that will not work with your product under any circumstances.

Software as a Service (SaaS) is relatively mature, and the phrase’s use predates that of cloud computing. Cloud applications allow the cloud to be leveraged for software architecture, reducing the burdens of maintenance, support, and operations by having the application run on computers belonging to the vendor. GMail and Salesforce are among examples of SaaS run as clouds, but not all SaaS has to be based in cloud computing.

See more about cloud computing/SaaS definitions at our Cleaning Up a SaaS Terminology Mess
post.

The packet analyzer is computer software, and sometimes hardware, that can intercept and also log digital network traffic. As data streams flow through the network, the packet analyzer captures each information packet and will decode and analyze the content (according to the appropriate RFC or other specifications). Packet analyzers are sometimes referred to as network analyzers, protocol analyzers, and sniffers.

Packet sniffers are versatile applications. For network maintenance, you can use packet sniffers to monitor network usage, gather and report network statistics, and debug client/ server communications and network protocol implementations. Security uses include the ability to analyze network problems and detect network intrusion attempt. In an offensive mode, packet sniffers allow you to gather information for effecting a network intrusion and spy on other network users. More controversial uses include the ability to collect sensitive information, such as passwords (depending on any content encryption methods which may be in use, and reverse engineer proprietary protocols used over the network.

Tcpdump is a popular packet analyzer with a command-line interface. It is used to capture and display TCP/IP packets (as well as other protocols) on the monitoring system’s network segment. This program is frequently used to troubleshoot network applications, but it can also be used to debug problems with the network itself, usually by detecting problems with the network routing configuration. Tcpdump can also be used to intercept network communications originating from another computer. By running tcpdump on a computer acting as a router or gateway, the user can display unencrypted information (such as that sent with TELNET or HTTP) including login IDs, passwords, URL requests, website content, and any other unencrypted data.

Wireshark, which was originally known as Ethereal, was renamed in May 2006 because of trademark issues. Wireshark is used for network troubleshooting, analysis, software and communications development and education. Even though Wireshark is similar to tcpdump, it has much more information sorting and filtering options as well as a graphical front end. The user is able to see all of the traffic that is being sent over the network, which is usually Ethernet, even though support is being added for other networks.

Ettercap is a network protocol analyzer and security auditing tool for Windows and UNIX. Ettercap can capture traffic, including passwords, on a network segment and it can be used to perform active eavesdropping. The software supports active and passive analysis of a number of common protocols, including encrypted protocols, and provides other network and host analysis features as well. Ettercap has four operating modes: 1) IP-based monitoring, in which packets are filtered by IP source and destination; 2) MAC-based monitoring, where packets are filtered by MAC address (this mode is useful for analyzing connections through a gateway); 3) ARP-based, which uses ARP poisoning to monitor a full-duplex switched LAN connection between two hosts; 4) PublicARP-based, which also uses ARP poisoning on a switched LAN, but is intended for monitoring half-duplex traffic between a victim host and other servers.

Red Hat announces the launch of the most recent edition of its benchmark open source operating system, RHEL (Red Hat Enterprise Linux) 5.4, with plenty of features to admire in this latest commercial Linux. A big plus is that virtualization is built right into this system. This RHEL version integrates KVM, Kernel-based Virtual Machine, which - in contrast to VMWare or XenServer - is built right into the operating system rather than as an add-on format. In fact, just as the name implies, it is incorporated directly within the Linux kernel.

Red Hats’ KVM can run up to sixteen virtual machines having up to 256 GB of RAM each. Because they function like a real machine would, you can use the normal RHEL management tools to make them work. Paravirtualized disk and network drivers have been incorporated into the LREL 5.4 for even better I/O performance. Even though Xen based virtualization is completely supported, the KVM hypervisor will only function correctly when used with the regular/non-Xen kernel.

RHEL’s also improved SystemTap toolset for performance monitoring will make it even easier for you to keep track of your C++ applications. With SystemTap, you can write instrumentation for a live running kernel with its simple command line interface and scripting language. A bigger library of internal “tapset” scripts, and newly published samples, are also available to assist with reuse and abstraction.

When you think of “the cloud” in IT, what benefits to your organization come to mind? With this “outsourced virtualization”, it certainly makes IT operations greener and prevents bottlenecks; it also helps organizations concentrate less on IT operations and more on their core functions. Analysts agree that going “to the cloud” is the first step towards the IT infrastructure of the future. Compared to traditional operations, moving your IT infrastructure into “the cloud” helps you deliver computing power within your organization in a much more efficient way, both financially and logistically - not to mention the benefits of not being responsible for upgrading outdated hardware and software on a regular basis.

Previously we talked about Cloud/SaaS apps advantages for clients, but what Cloud will mean regarding systems management and monitoring is as follows: Small and mid size businesses are looking for two sorts of services regarding systems management: performance and external end to end uptime, led by such companies as Gomez and Keynote; and network monitoring and servers, minus the big four with big enterprise focus-for this space, clients usually think of Adventnet (now Zoho), Solarwinds, and open sources vendors such as Zabbix, Nagios, and Zenoss. Though the first category is natural for Cloud apps since they have to be out of a corporate firewall, the second one is usually conventional software.

Our customers have told us that they have two primary needs. First, they need easy and quick way to monitor their networks and servers through a centralized application utilizing all benefits of multiple tenant architecture, benefits we’ve described in past posts. Some vendors generate an instance of each application from the cloud infrastructure, from Amazon or some other vendor. However, these are not true cloud applications, comparable to those operated by Salesforce.com. IT engineers and managers will benefit from faster deployment, lower initial capital requirements, and easily scalable architectures that will allow them to grow a little bit at a time.

Secondly, internal servers and network monitoring needs to be merged with external end-to-end monitoring. This should be done in a single product which delivers a consolidated view of overall system performance. External monitoring is focused on business transactions and end-user experience tied to service level agreements provided to third parties whilst IT Staff were responsible for the internal functions of server and network monitoring. The role of IT in business is gaining more importance with time as IT goals become more aligned with overall business goals.

System administrators may wonder what benefits they will reap from a changeover to centralized systems monitoring in the cloud. We tell them they will have better speed of deployment and more control and visibility. They can take advantage of proactive monitoring and early notification. Some technologies and administrators would have liked but couldn’t have considered because of high cost and complexity will now be available. It is safe to say that the control over end-user experience administrators will have with consolidated monitoring will ease a lot of their workload. In short, they can put their efforts where their skills are best used and gain recognition and admiration from their supervisors.

Orders, not errors, are what you want from your new ecommerce website. Keeping customers happy and product/service requests rolling in is easy to do with Synthetic Transaction Monitoring Services. With Monitis Transaction Monitoring in the Cloud, there can be constant monitoring done on simulated transactions on websites. It’s ideally used for ecommerce enabled sites and for online banks and other service providers.

With Monitis, you keep track of the usual steps (search, purchase, order tracking, etc.) that users take to finish transactions on your site. Monitis can also assist in writing scenarios free of charge. The steps can then be simulated regularly. If these steps are not finished, which means there’s an issue, then you or the support staff will be informed right away and the corrections can be done. This really helps if the site is using third party services for ordering processes, like processing credit cards or doing real time shipping calculations.

For the past fifteen years, Bessemer Venture Partners has focused its investments in the Software as Service business model, and has investments in over a dozen projects such as  Verisign, Cyota, Postini and Trigo. The emergence of software-as-a-service, we at Bessemer Venture Partners believe, is the single most important trend in the modern software industry. This monumental shift, from software as product to software as service is just the beginning. Bessemer’s top ten laws for being “Saas-y” was developed in early 2008 to describe and incorporate the best practices and counter intuitive discoveries that Bessemer has discovered in the SaaS field.

1. Your most important monthly financial measurements are Churn, Cash Flow, and CMRR (Committed Monthly Recurring Revenue). For years, software executives have been taught that Bookings are the most important single measurement of a growth software business. That is not the case for SaaS companies, where Bookings data is highly misleading. They instead focus on Monthly Recurring Revenue, which is simply the total amount of all current recurring income from subscriptions. BVP believes that a forward view of Committed MRR is even more important than the present MRR data.

SaaS executives need to track churn in detail from a “logos lost” (lost customers) perspective as well as the amount of lost CMRR. It’s very difficult and expensive to grow subscription businesses if you have moderate customer churn, and prohibitive if your churn is high. Whereas the largest legacy enterprise software companies literally made hundreds of millions of dollars over the last decade with “shelf-ware” projects that never got fully implemented, project failure is not an option for SaaS businesses or the customer will simply turn you off, regardless of your contract terms. The top performing SaaS companies typically achieve annual customer renewal rates above 90% - with most of the churn due to death (bankruptcies) or marriage (acquisitions) - and over 100% renewals on a dollar value basis due to up-sells into this installed base.

Cash Flow is the other key metric. To be fair, visibility into the current cash position and the change in the cash position has always been important for software executives, but is even more critical for SaaS businesses because the working capital requirements are higher and the payment terms are often stretched out over the term of the contract. Given the high cost of capital for private SaaS companies, wise executives will often offer slight MRR discounts to customers in exchange for quarterly or annual pre-payment terms, and provide incentives for their sales force accordingly.

2. Customer Acquisition Cost (CAC) and Customer LifeTime Value (CLTV) are the best indicators of long term value creation. It can be argued persuasively that SaaS is a lousy business model because your costs are front-loaded and your revenue only arrives in modest monthly or annual payments. However, as we know from the cable industry, subscription businesses can be very profitable over time. The key to long term financial health is to keep customers happy so their payments keep coming, and coming, and coming…and over time add up to some really large numbers.”

3. The Sales Learning Curve is critical for any SaaS business to climb, and it must be done before you grow your sales force. Stop at three sales representatives until at least two of them are making at least $100K MRR quotas. Often, software businesses grow their sales force before they have a clear model for how the product should be sold. This problem is especially acute for SaaS businesses because of the large upfront investments in IT infrastructure necessary for the service to even be provided. Those precious cash reserves must be carefully husbanded, and not burnt through by a bloated sales department. When several sales reps are able to regularly sign contract values equal to twice their fully burdened cost of sales that is when you are ready to expand your sales force.

4. Both finding new customers and retaining old ones are critical for the health of the business. As soon as you have a customer base large enough to service, add to your sales force (”hunters“) with account managers (”farmers“) dedicated to customer retention. Both sales and account management should be paid on the basis of CMRR growth. This is especially important once a SaaS company reaches the sales inflection point.

The level of sales and marketing investment should be computed using the CAC Ratio, which can be computed with a simple look at your GAAP profit and loss statement. Simply divide your annualized gross margin added over the last three months by sales and marketing costs incurred over that period. A CAC Ratio above 1.0 indicates that you need to invest more money in sales and marketing because your customers are making themselves profit centers within a year. (BTW Monitis CAC is above 1)

You should also pay attention to Customer LifeTime Value, or CLTV, which is simply the present value of the profit stream for a customer, minus the cost of acquiring the customer. Younger companies have little track record and hence can only guess at a customer’s lifetime. BVP uses estimates of 3-4 years of SMB customers, while enterprise customers should stay for 5-7 years. The “Five C’s of SaaS Finance” are thus CMRR, Cash Flow, CAC, Churn, and CLTV.

5. Traditional IT channels, where one tries to sell the computer service to the head of the IT department do not work in a SaaS model, for the simple reason that a SaaS model makes redundant much traditional IT work, and IT managers are reluctant to manage their departments into superfluity. Instead, focus your efforts in business development towards business services channels. This will require additional work, as these new sets of partners will not be as comfortable with the technical aspects of the product, but will understand the power of a SaaS product while many ISVs and SIs will not. Focus on the marketing department, the payroll department, and businesses that need computer services but are not yet large enough to have dedicated IT departments.

6. By definition, your sales prospects are online - Savvy online marketing is a core competence (sometimes the only one) of every successful SaaS business. You sell a product that requires an internet connection and a web browser for access, which means your prospects are online! Numerous studies show that your customers are now doing most of their primary research online, and it should not surprise you. This is a clear example where business-to-business (B2B) marketers need to learn from their business-to-consumer (B2C) counterparts. The most innovative B2C companies are lead generation machines, leveraging search engine optimization (SEO), viral marketing, Search Engine Marketing (SEM), email marketing, and other technically-advanced methods. Yet many B2B companies don’t have a clue.”

7. Globalization obstacles hinder SaaS vendors more so than traditional software companies, which is why you should establish yourself in North America first. After you reach $1 million in CMRR, then research hiring European sales and services executives pushing customer demand. Only consider utilizing Asia for post-IPO. Saas companies consistently address concerns about latency, data access and security by replicated local datacenters, in-country customer support personnel, packaged integration and other regional software and SaaS products due to different structure and service goals throughout the industry (Monitis doesn’t follow this rule although).

8. You should have one and only one version of the code in use. That means you should have one instance of the program, one datacenter for all your tenants, and no on-premise deployment at all. The best SaaS companies follow this advice. Multiple instances and single-tenant offerings are used by legacy software businesses whose architecture cannot be redesigned on the fly, in the way that SaaS companies can. Virtualization can allow you to make multiple instances of your product, but the additional engineering complexity will raise costs and reduce ability to stay nimble. Your SaaS product should always begin by being multi-tenant and single instance. (we also think it is very critical)

9. “Service” is the most important part of a Software-as-service business, not “software”. Detailed usage data is produced by your customers every time they use your service. Use it! Product managers in a licensed software model will spend hundreds of man hours and hundreds of thousands of dollars to just create a vague picture of how the customer uses their product, without ever really knowing how the product was used every day. With a SaaS model, that data is instantly available. Use that data, analyze it, and incorporate what you have learned to evolve your product each and every day into something that the better serves the customer. Companies like Apple, Google, and Facebook have already incorporated the use of usage data into their business model and customer service strategy. Now, SaaS companies can use the same methodologies those companies use on the business of software itself.

Three different levels of Service Savvy exist, and a SaaS business should try to achieve them all. The first level requires basic monitoring to see if customers are likely to churn or can be sold additional products. This competence is required for any SaaS business, and shouldn’t be hard to achieve because you know how and how often your customers use your products. Simply track key usage metrics and determine which customers are getting the best value for their money (these are your up-sell candidates) and which ones aren’t (these may leave unless you intervene). You need to cooperate with your marketers to create internal reports about customers with low usage, and should send emails to customers whose behavior changes so that you can figure out why they have altered their usage of your products.

Second, expect to test and develop your software quickly based on feedback from your customers. Analyze the ways in which customers actually use your product instead of how they say they use it. Certain pages, features, and components get more use than others; find out why and whether they can be used to create more opportunities. Third, analyze all of the data that you collect to get benchmarks for your customers and determine best practices; you can actually help your customers’ businesses grow by helping them use your product more effectively, while your business can learn from the practices of your best customers.

10. Recognize that you’ll need a lot of capital to get started; you have to spend a lot on sales personnel and R&D, and you cannot reasonably expect to make a consistent profit for at least four years. Therefore, be ready for the long haul, and don’t have too many frills in your budget. Many SaaS companies tried to expand too fast, and wound up going out of business after promising starts. A sizable cash cushion should be baked into every operating budget, and faster growth will actually result in more cash burned. BVP believes that second-generation SaaS companies may become more efficient than those in the first generation, but they’ll still need a lot of capital to have a good chance of becoming successful SaaS businesses.

Some rules are meant to be broken, but not all rules are meant to be thrown out the window, you have to be choosy about which rules you break at which times. These laws will help you run your SaaS business better. But even though the best companies are constantly pioneering through their fields, but BVP thinks you can excel even if you decide to disregard one or two of these rules. Several companies BVP has worked with prove that you can break one of these laws and still succeed. But if you are looking at these Ten Laws and questioning more than just one or two of them, you should stop and really examine your business.

There is a new announcement from Amazon Web Services (AWS) which enables us a secure and seamless connection between available IT infrastructure and the AWS cloud. It is the Amazon Virtual Private Cloud (VPC) which is going to provide us these facilities. The existing securing infrastructures like firewalls, security services can be extended to our AWS resources by using Amazon VPC which bridges the IT infrastructure and the AWS resources through a Virtual Private Network (VPN). The integration of the Amazon EC2 computer resources is now possible using Amazon VPC. This can be extended to the integration of the Amazon VPC and AWS services in the future.