|
|
article page | 1 | 2 | 3 | 4
and the Community Communication Authority (CCA) have teamed up to bring Wireless Broadband (utility-grade broadband) to rural America. This model of starting with municipal security services and expanding to become local “last mile” connectivity networks for homes and businesses is starting to catch hold. Many small companies and cities are actively watching these initial business trials and seeking to duplicate positive results.
Management models
Will these devices be managed locally or remotely? Local management makes sense if management services can become automatic, robust and simple enough to be deployed into home and small business environments. Management services today are not developed with this degree of autonomy and simplicity. Architecturally, the closest local management model today is the remotely deployed management agents used in large IT software distribution networks such as IBM’s Tivoli. This model can meet scaling needs if agents can become autonomic, remotely deployed and simple.
Alternatively, the management model could be remote management like the device managers for mobile phones from service providers. A model like remote management agents implies a local management network in the home or small business with agents communicating back to the larger network when necessary. On the other hand, Service Providers today can monitor phones via presence on the cell network pushing down configuration information and can update phone software as new services become available.
We believe that current phone management models offer the most promise going forward. However it is not yet clear what is needed so this approach can handle a hundred to a thousand times more devices.
|
|
optimize consumer preference realizations and increase efficiency of consumers and businesses. Some of these transactions may be quite small, and there is a cost to performing the query and getting a response – so micro transactions (those of pennies or less) will become commonplace. Traffic and transactions from these IMS devices will greatly exceed the scope of existing credit card networks.
It is unclear that IMS software vendors are planning for this degree of acceptance. Both the IMS software services and the management of those services will need to handle enormous transactions. And the security of these transactional services must be absolute.
Given all these near future IMS and SOA enables services, and all those home and business devices connected to the net, even more backbone bandwidth is actually needed. Major suppliers will be moving to full 40 Gigabit build outs. With IMS and SOA services, we will be talking of “billions and billions” of connection streams riding over these trunks at any time. These services will connect with vast network resident computing grids.
Get with the grid
Today service providers see grids only as opportunities to sell managed services to big businesses using the service provider product models based on frame relay management service products. This narrow viewpoint may mean service providers get left behind – subsumed by grid savvy companies like Sun and IBM. Service providers should immediately advance their models of how to manage grids and begin to put grids into the top of their strategic thinking. Transaction networks for micro transactions will need massive grids. Management applications will require the power only supplied by distributed
|
|
|
|
While the communication networks are evolving so to is software. Service Oriented Architecture (SOA) will allow everything (including software, devices, connections, etc.) to have functional and management services it connects with. These services need not be local, and can be global in distance and in scope. With adequate bandwidth, network resident, remote virtual services will become one of the largest consumers of QoS connectivity as they are used to record lives, provide intelligent personal assistants, manage all financial transactions, and oversee all the things in ones life – rather important things like kids, homes and cars.
IMS is the service providers’ bet for providing and managing QoS bandwidth. But less well understood is that IMS likely will become the network side of the SOA software for service connectivity. By providing guaranteed, managed QoS connections, people will feel comfortable enough to offload the support services and management to network resident services, which releases some stress for remote management systems, but requires management of the connectivity of linking with these remote network services.
Also IMS enabled presence and location information will allow people to form temporary data service connections with all the business and people they pass. With presence and location, from the perspective of service connections, the last mile drops to 100 feet. As people walk past businesses, their IMS phones (if we can still call this advanced device a phone) will query stores and restaurants for the service preferences of the phone’s owner establishing transactions and setting phone owner schedules to
|
|
grid to process and control these loads. IMS itself needs to become a distributed grid service.
Management services
So far there is no management scaling problem because FTTP and IMS service is rolling out so slowly, a few hundred thousand homes or SMBs at a time. But what happens when 30 million homes are connected and each home is managing 5 services and 30 devices? And when it is 90 million homes with 200 services in each? Or when 300 million USA residents will have connectivity to a constantly changing environment of networked things and services? And certainly twenty years from now when Indo-China passes the USA in communications demands; and later if Africa and South America need these as well?
So how do you manage QoS on several billion connections that make up your part of this global network? How do we get 100 billion connections to inter work reliably, at the same time? Again it is clear that today’s approaches to management cannot support a communications infrastructure of this scale, scope, and complexity. It is unlikely that any centralized management architecture can. But the good news is that this time, “soon” gives us perhaps a decade to mature new OSS systems. In this time, if the infrastructure and OSS communities accept this vision and accept this precipitous horizon of need, new systems will arise.
The seeds are here today in experiments with distributed grid software and autonomic agent systems. Better to get behind growing these technologies into mature markets, than to expect another bubble bursting to bring these trends to a standstill.
|
|
|
article page | 1 | 2 | 3 | 4
|
|
|
