Opportunity Analysis Project UNICOM "Where do you want to compute today?" ----------------- Executive Summary ----------------- "There are a number of really strong indications that, regardless of the fate of Sun and the SunRay product, the world is moving in this direction and the next wave of computing will be based on SunRay-like technology." -- Duane Northcutt, Sun Microsystems For computer users who need a maintenance-free, secure, and mobile computing environment, UNICOM is a computing utility that provides high-speed infrastructure for global access to network computing systems. Unlike those who provide distributed computing resources, our product is easy to use and offers a complete personal computing experience regardless of location. With the exponential growth in network bandwidth and computing power, it is now possible to push computing away from the edges and to centralize the management of computing resources. UNICOM aims to provide a computing utility in much the same way that there are power utilities. People will pay for computing resources such as processor time, memory, and disk space on an usage basis. The access to the UNICOM system will be through very simple thin clients that will not need to ever be upgraded or maintained. UNICOM intends to initially target the small business market, as small businesses have IT needs that UNICOM can provide, while not having a large budget for their own large and expensive IT department. The market looks favorable at this time. There are many possible competitiors in this area, but there are also many possible partners. By developing partnerships and quickly building a large user base, UNICOM will be able to compete with larger businesses if they decide to enter this market. ----------------- Concepts & Trends ----------------- The computer industry continues to see a trend of exponential growth in technological performance. Moore's law postulated by Gordon Moore in 1965 says that the density of transistors on integrated circuits doubles every 18 months. This has held true for over 30 years and should continue for many more years. Since price depends heavily on the size of the chip, Moore's law implies that for the same size hardware, one can make faster and better processors, memory, and storage, without increasing the price. In addition, network bandwidth has recently shown explosive growth. This trend has made it feasible for the first time to implement the high-speed, high-capacity provision of distributed computing environments. The number of computer users also continues to grow. Many of the new and old users alike have no desire whatsoever to own or maintain their own computer. There is also a second group of users who require computing mobility. They may need access to their files while on a business trip. The current best situation is to carry a laptop around, which is inconvenient and also prone to theft or physical damage. A third group of computer users are those who want access to a lot of computing resources. With those who do own computers today, the amount of wasted computing resources in terms of processor, disk, and memory usage is only growing larger as a result of Moore's Law. UNICOM addresses all three of these groups to provide a maintenance-free, easy-to-use, secure, personalized and mobile computing environment. We envision UNICOM as a computing utility in a similar way that there are currently electric and power utilities. Similar to how power plants provide electricity on demand, UNICOM will be a computing plant that provides computing resources on demand. No longer will it be necessary to have one computer per user, but rather we can view a network of computers as one large machine that will be shared by many users. Users will access the UNICOM system through simple thin-clients using smart cards or other forms of identification that can be easily carried around. Duane Northcutt, who led the development of the SunRay product at Sun Microsystems, agrees that UNICOM's vision is the right direction to go. "The service model is clearly the direction in which the world is moving and I think it will be the dominant model in the post-PC era." Duane believes that as bandwidth becomes more plentiful and more important, the compute/communicate tradeoff will move the computation away from the "periphery of the network and back into the infrastructure." ------------------- Potential Customers ------------------- Initially, UNICOM plans to target the small-to-medium-sized business market. The targeted businesses typically have fewer than 100 employees. Over 5 million of these businesses exist in the U.S. alone and they frequently operate on limited information technology budgets. They would thus view a competitively priced, centrally maintained computing service as very attractive. UNICOM addresses core needs of its customers. First, given continued increases in IT spending (approx. 5-10% annually) and the ongoing shortage of qualified IT personnel, small-to-medium-sized companies would appreciate the constant hardware and software maintenance and upgrades and guaranteed system uptime UNICOM provides. The UNICOM service would be an alternative to the IT personnel currently employed or contracted by these businesses. Using UNICOM would also provide mobility for those business people who need to travel and have access to files. UNICOM's money- and time-saving characteristics would be a boon to small businesses. ------------------ Market Environment ------------------ UNICOM foresees an addressable market of 5,700,000 small-to-medium-size businesses, each with 100 or fewer computer-using employees (source: Small Business Administration). The large existing market for information technology (IT) continues to grow, with approximately $2.5 trillion globally in 2001, and 200 billion for IT expenditures by small businesses. Also, large and rapidly growing markets exist for IT services and application outsourcing. These businesses would presumably use one thin client per computer-using employee. Our revenue model will be very similar to that of utilities. People will only pay for software, memory, and processor capacity that they actually use. While today, people buy lots of bundled software with a computer that they don't use, this model should save money for users, and encourage them to move to using the UNICOM service. For software developers and vendors, we will give them rights to install their software and to set the price to charge users who use that software. We then take a piece of this fee for our own revenue. UNICOM foresees the following per-user costs: THIN CLIENT $300 one-time cost NETWORK CONNECTIVITY $20/month This assumes that UNICOM purchases bandwidth at a cost of $200 per month per Mbit/sec, and that ten users can be served per Mbit/sec of bandwidth -- i.e. each user receives a sustained 100 Kbps of bandwidth, or approximately the speed of a DSL connection. CENTRALIZED COMPUTERS $35/month This assumes that UNICOM uses $1000 computers, that each computer can serve approximately five users, and that each computer needs to be replaced every six months. IT PERSONNEL $5/month This assumes that UNICOM hires IT staff members with salaries of $60,000 annually, and that each staff member is able to support 1000 users (200 computers). Thus, it will cost a total of $20 + $35 + $5 = $60/month to support the typical user. UNICOM anticipates the following per-user prices, based on the utility pricing model (seen in telephone service, for example): BASIC SERVICE $5.00/month STORAGE (per GB) $1.00/month MEMORY (per MB) $0.02/month (includes CPU) NETWORK TRAFFIC (per Gb) $0.15/month BUSINESS SOFTWARE PACKAGE $30.00/month (variety of packages available) UNICOM expects the typical user to spend approximately $100/month for the service. Current business IT expenditures are currently in the range of $200 per employee per month, so the UNICOM solution can offer substantial savings. The projected growth of UNICOM's user base and revenue appears below. This is based on an average 25 computer users per small business customer. YEAR NUMBER OF BUSINESSES NUMBER OF USERS REVENUE 2001 1000 25,000 $ 2,500,000 2002 10000 250,000 $ 25,000,000 2003 50000 1,250,000 $125,000,000 2004 100000 2,500,000 $250,000,000 In addition to customer charges, UNICOM also sees an opportunity to gain additional revenue from advertisers by displaying advertisements on thin client displays when they are not actively being used. UNICOM also sees the opportunity to eventually expand its service into the large business market, as well as the consumer market. ----------- Competition ----------- UNICOM faces competition on several fronts. This competition comes from hardware providers, software providers, and service providers. In the hardware arena, potential competitors include Sun Microsystems and Oracle's New Internet Computer Company. Sun's SunRay family of enterprise-class information appliances operate as thin clients, with processing and storage taking place at a centrally located server. However, these products are intended for sale to the enterprise -- the responsibility for their operation and maintenance rests with the enterprise, not with Sun. Similarly, Oracle's "Network Computer" vision, along with its spinout New Internet Computer Company, anticipate a world of thin client information appliances with minimal capabilities, accessing richer services and information sources over the network. These information appliances are not unlike those UNICOM intends to distribute to its customers. However, UNICOM's focus is on providing a service -- not on developing hardware. Amongst software companies, UNICOM faces competition from Centrata, Entropia, and United Devices (as well as non-commercial efforts Seti@Home, Folding@Home, and distributed.net). Microsoft is an additional potential competitor. Microsoft's recent .NET initiative is intended as a platform for network information and communication services. This highly distributed design is in contrast with UNICOM's vision of a centralized source and repository of such services. Centrata, Entropia, and United Devices are providers of distributed computing resources for enterprise applications. These companies develop software that allows for the assembly of networks of personal computers. They then provide a service component, selling network processing time to corporations. However, their distributed architecture is in contrast to UNICOM's centralized model. The @HOME technologies, non-corporate in nature, use a distributed network of personal computers to perform data-intensive operations, allowing for supercomputer-class processing power. Similarly, distributed.net uses a distributed network of personal computers to solve time-intensive operations such as the breaking of cryptographic ciphers. Service-based competitors include TAOS, "The Sys Admin Company," and companies with similar offerings. TAOS provides outsourcing services for business customers. TAOS staff members handle the setup and maintenance of IT systems for those customers. UNICOM also faces competition from easyEverything and other Internet cafes. These cafes frequently make use of a "utility" pricing model not unlike UNICOM's. However, users of easyEverything must be physically present at one of the cafe locations, and must make use of a full-fledged computer (rather than a simpler thin client). An additional service competitor is Scale8 (and its non-commercial cousin, Berkeley's OceanStore project). Scale8 is a storage service provider that gives the enterprise access to a "global filesystem" of its corporate data -- all data can be accessed as if it were on a local device, regardless of its actual location. UNICOM anticipates building partnerships with many of these potential competitors. Because UNICOM will need hardware and software suppliers in order to set up its own systems, companies such as Sun will be well-positioned to deliver the necessary technology. According to discussions with the head of SunRay development, Sun already engages in these types of partnerships. The SunRay was meant to be used for the delivery of network-based services, and it was meant to be used to support service providers. As Duane Northcutt summarized, "We were looking beyond ASPs to TSPs (Total Service Providers) or FSPs (Full Service Providers) -- i.e., companies that offer services that include not only the backend software, but the infrastructure and access devices." Duane also gave several real examples of partnerships that Sun had with service providers. This is exactly the model that UNICOM is looking at, so a potential partnership with Sun would be beneficial to both parties. --------------- Success Factors --------------- UNICOM follows a business model similar to those of utility providers and telephone companies. Users pay a basic monthly rate for access to UNICOM resources, as well as additional fees based on their usage of those resources (CPU, memory, storage, and network capacity). UNICOM's success is dependent on customer acceptance of the computing utility concept. That acceptance will require a paradigm shift in ways of thinking about desktop computing. To spur on acceptance, UNICOM may engage in special promotional tactics, including trading customer computing equipment in exchange for thin client devices or giving away thin client devices in exchange for service contracts for a set period of time (similar to the cellular telephone model). A not small risk is what Duane Northcutt called "intellectual inertia." The SunRay and the ideas behind it are a "disruptive technology" and getting people to accept it is no easy task. People in general are not comfortable with technology change. Customers who use the SunRay product today are universally positive about it, says Duane. The main problem is that many people are tied to the Windows platform and the applications on it, so the SunRay is mainly used for web-based and some UNIX-based applications. While UNICOM intends on supporting many platforms including Windows, we will also have the problem of intellecutal inertia. Companies already are managing with what they have today, and may not see the need or benefits of change. There may be environments for which UNICOM is not appropriate. For example, at a large institution such as Stanford, distributed computing resources for the campus is an important priority. There are 35,000 users being supported on Stanford's Leland Systems. The workstations cost about $12,000 each, with a lifetime per workstation of 3 years. Each workstation can have about 15 users at a time before being too overloaded. For staff to maintain the systems, there are 11 full time staff with an average salary and benefits of around $90,000. This means that it costs about $300/year to support one user. This doesn't include many of the other support structure required in this distributed environment, but even including these costs, we can assume that it won't approach $100/month. Dwayne Virnau says that for the large distributed environment at Stanford, "the setup works extremely well." They have specialized services so is not clear how well our model would work. Large businesses may be in a similar position as Stanford, with an established system that works well. By focusing on small businesses, it is more likely that UNICOM will provide a valuable service to them. Although Dwayne says that, "at Stanford, we've solved our system administration problem," he does admit that "what isn't maintained are desktop machines." And as he summed up, there "has to be a market there...make money off selling system administration." UNICOM achieves profitability by charging each user more than it costs to service that user. A large user base is necessary in reaching this goal because of the high costs associated with the initial deployment of UNICOM's computing infrastructure. The success of the company is dependent on the ability to attract and retain this sizable customer base. UNICOM's competitive advantage requires rapid customer acquisition and first-mover advantage. Customer retention will be possible due to the high switching costs that would be associated with moving from one computing utility provider to another, particularly since corporate data would be stored within the UNICOM system and could not be easily transferred to another system. UNICOM's success is also dependent on the continued growth of the market for information technology. According to recent reports from Investor's Business Daily, IT spending has slowed, though it continues to grow at rapid rates. While the success of an opportunity like UNICOM is dependent on many factors, the trends and market analysis have shown that this is the direction that computing technology is moving. The potential impact of a system like UNICOM is enormous, and is an opportunity that cannot be overlooked.