Showing posts with label Savvis. Show all posts
Showing posts with label Savvis. Show all posts

Friday, October 5, 2007

Level 3 Provisioning Issues?


Level 3 Communications has made seven acquistions in the last 18 months, including Broadwing, part of the Savvis portfolio, TelCove, Looking Glass and ICG. Several of the buys related to its new content delivery network portfolio. But many of the sizable acquisitions are related to its core bandwidth business.

If you have been around the business long enough, you know what is happening in the back office. Disparate systems are running in parallel. Manual reports have to be built. Billing systems don't talk to each other. Inventory cannot be interrogated in real time. Provisioning backlogs are the inevitable result.

That is one possible reason Level 3 reported relatively light revenue and earnings growth in the most recent quarter. Demand for its services isn't the problem. If anything, in light of the back office issues, demand possibly is outstripping provisioning.

Any company would have at least some issues getting new customers provisioned efficiently were it to digest as many acquisitions as Level 3 has made recently. So we would not be surprised if the company is having issues getting the new customers onto the network.

In fact, it probably is safe to say that bandwidth in service has grown at an unprecedented level over the last year. If one looks at total in-service bandwidth provisioned by Level 3 in its entire history up to about 12 months ago, the amount of bandwidth in service probably has doubled again in the last 12 months.

That rate of growth would cause issues for any service provider. So we wouldn't be surprised if there was some hiccup in provisioning new orders. In fact, the backlog could be great enough that three months might have to pass before the provisioning teams can catch up. Level 3 wouldn't be the first company to have problems with too much success.

"Tokens" are the New "FLOPS," "MIPS" or "Gbps"

Modern computing has some virtually-universal reference metrics. For Gemini 1.5 and other large language models, tokens are a basic measure...