More Service Problems Plague @Home1/17/1999 7:00 PM Eastern
High-speed Internet-access company @Home Network once again
found itself confronting a blizzard of bad publicity last week, this time pertaining to
service glitches in Connecticut and Rhode Island.
By week's end, the company was to have taken the
necessary steps to overcome the impediments to high-speed-data delivery that triggered an
outcry from unhappy customers in the region. But officials acknowledged that similar
problems could crop up elsewhere, as pressure on capacity -- generated by ever more users
consuming ever more bandwidth -- continues to mount.
"We just plain screwed up," @Home chief
technology officer Milo Medin said, in reference to the problem in New England. "We
put an order in to Sprint [Corp.] and TCG [Teleport Communications Group] to upgrade
capacity in that market, and we didn't get it done on time."
@Home's nationwide backbone connects with the cable
systems of its Connecticut and Rhode Island affiliates -- Tele-Communications Inc., Cox
Communications Inc. and Cablevision Systems Corp. -- through a single regional operations
center in Hartford, Conn., Medin said.
Those connections were in the process of being upgraded to
DS-3 (45 megabits per second) capacity from DS-1 (1.5 mbps) at press time.
@Home has been pursuing a "huge number" of
similar upgrades since the third quarter of last year, Medin added.
"I think that there are areas where we've done a
good job of staying ahead, some areas where we're running close to the edge and areas
where we didn't get new capacity installed in time," he said. In areas that fit
the last description, he added, the company is "seeing problems like Hartford come
In early December, @Home had a run of bad luck in Fremont,
Calif. -- a TCI market where a concurrence of unrelated problems caused massive slowdowns
in access rates and an outpouring of bad publicity. While the problems in the two regions
differed, the common thread between Fremont and New England was @Home's lack of
responsiveness to customer complaints until the situation generated an uproar.
Customers in southern New England reported that they had
been noticing an ever greater deterioration in their access speeds over the past four
months, to the point where by last week, many were receiving data at a crawl -- well below
normal dial-up analog rates.
Adding to their outrage, they said, was a general lack of
acknowledgement of the problem and, in some instances, misleading and even rude responses
from customer-care personnel.
"I ran a comparison of downloads to computers using
the cable modem and a 28.8 [kilobit-per-second dial-up] modem side-by-side," said
West Hartford, Conn., resident Steve Greczkowski, a management-information-services
director at a law firm in Hartford and organizer of a user group for people having trouble
with the service.
"Web pages that took three seconds to download on the
dial-up modem were taking 30 seconds to one minute to download over the cable modem,"
Greczkowski, a TCI customer, and others sending e-mail to
the news group reported that customer-care personnel suggested a variety of possible
causes for their problems -- such as poor electrical connections to the modems, or the
effects of cold weather -- without acknowledging the fact that there was a systemic issue.
"It wasn't until we started getting vocal with
the news group that they became responsive," Greczkowski added.
"Some of us are seriously considering entering a
class-action suit against @Home," said Manchester, Conn., resident Chris Dupuy, a
senior systems engineer at Phoenix Home Life Insurance Co., who is served by Cox.
Dupuy was especially concerned over the possibility that
@Home might impose an upstream-speed cap of 128 kbps per user, as it has done elsewhere,
although @Home contended that such a cap squares with the stipulations in subscriber
contracts that mandate nonbusiness applications of upstream capacity.
But few among senders of e-mail reviewed by Multichannel
News or those interviewed indicated that they were terminating service.
"I intend to stick it out, because when it works,
it's great," Dupuy said.
A common thread of complaints from @Home customers
nationwide experiencing various problems was that, as Greczkowski put it, "The left
hand doesn't seem to know what the right hand is doing" -- a reference to the
divisions of help-desk responsibility between the cable operator and @Home.
For example, Greczkowski and other customers in the
Hartford region said they were told by TCI's help desk in Denver that their most
recent slowdowns were the result of a malfunctioning router. Greczkowski said he was given
that explanation by a senior staffer at TCI last week, who began reading a message aloud
that had been posted the night before from @Home, only to find, as he read, that the
problem was the DS-1 link.
"He slowed down as he read the message, and then said,
'I should have had my ducks lined up before I called,'" Greczkowski said.
Similarly, an @Home customer in Kenner, La., a suburb of
New Orleans, posted an e-mail recently detailing several instances of service failure.
Each time, he was put through several procedures on the assumption that the problem was
his alone. In fact, the problem invariably was a faulty node in the newly installed data
Jerry Lenaz, high-speed-data product manager for the Cox
system in New Orleans, said the miscommunication resulted from the fact that tier-2
customer-service support was handled by @Home, which was only responsible for knowing
about the data-network performance through the backbone to the headend.
The Cox system is responsible for monitoring node
performance, which means that Cox's tier-1 support group in San Diego would be the
conduit for passing information about a faulty node to the @Home group, which apparently
"We probably could have done better," Lenaz said.
Compounding the complexities of communicating problem spots
from different points of responsibility for network performance is the fact that
@Home's role in customer support varies from one MSO to the next. Tier-2 calls --
calls that need responses from technical staff -- are handled by the MSO in some instances
and by @Home in others.
"We're a little more flexible about customer-care
responsibilities than is generally known," Medin said.
Along with regional data center-to-headend capacity
problems, @Home is faced with other issues, including what it sees as violations of the
restrictions on upstream usage.
"People agree to use this as a residential service
and, in a few cases, they try to run business applications," said Dee Finley, a
spokeswoman for TCI's Connecticut operation.
Finley said @Home can identify these people, and it is
taking action to prevent abuses. In addition, the company is drawing up a new service
agreement that is more explicit about the prohibitions, she added.
There are other problems, as well. For example, Finley
said, one user's modem -- apparently connected to a server on-premises -- was acting
like a router, terminating packet flows from other users and leaving them without useful
This problem also contributed to the difficulties in
"@Home assures us that they can and will monitor the
network for such problems," Finley said.
One of the more mystifying issues was packet losses widely
reported by users who conduct "ping" tests, which put packets out on a
round-trip over the network to determine overall performance. E-mail from users in
Southern California, Toronto and New England last week attested to packet losses ranging
from 22 percent to 38 percent.
Medin said such loss rates would represent a breakdown of
the network on a scale that has not happened.
"Normally, if there was an upstream problem of that
nature, it would be escalated to where I'd know about it," he said.
In general, the outcrop of problems is attributable to the
rapid growth of a still very young service, Medin noted. As @Home moves to its new
nationwide backbone later this year, using high-speed fiber lines and new routers, it will
have much more capacity and much greater control over network management, he added.
But in the meantime, the road bumps could get worse.
Not only are there surging bandwidth demands to be met as
@Home waits to complete the new backbone; there is also the challenge of managing the
transition from proprietary modem systems to the new standardized DOCSIS (Data Over Cable
Service Interface Specification) modems.
"We're using two proprietary systems in our New
England networks, and we are introducing DOCSIS in others there," said Scott
Hightower, product manager for voice and data at Cox New England. "It's only
going to get more complicated as we move to multiple brands of DOCSIS modems that are
distributed through retail outlets."