V. Batch Size, Capability of SBC's Processes to Meet Demand, and Provisioning Interval
The first task to be completed is the adoption of a "batch size." We address this task for SBC in this section.
In addition, the third task to be completed is to determine whether or to what extent SBC's existing processes, which handle hot cuts on a one-by-one basis can be scaled up to accommodate multiple hot cuts in batches sufficient in volume to meet the increased demand that would result from the elimination of UNE-P.
We address these two tasks together because the TRO called upon the states to determine the appropriate volume of loops to be included within the batch in order to meet demand assuming a 27-month transition period.15
Finally, a key to a seamless transition is the interval required to complete the migration. We address this aspect of Task 2 in this section.
A. Position of SBC
SBC's proposed batch processes are built upon its current one-at-a-time hot cut processes. SBC emphasizes that its existing processes are well established. The SBC Panel witnesses claim that SBC's existing manual operations and Operating Support Systems (OSS) capacity can meet current hot cut demand, and can easily be scaled to meet the increased hot cut demand that would result from elimination of UNE-P.16 SBC witness Chapman calculated the volumes of both BHC for the embedded base and ongoing hot cuts that would be required absent UNE-P switching. Chapman based her analysis on data for embedded UNE-P lines as of September 2003 and the highest monthly inward UNE-P activity previously experienced at each central office.17
SBC witness Hopfinger also claims that the FCC has determined that SBC's existing hot cut processes, are "scalable to meet reasonably foreseeable demand."18 The basis for the FCC's evaluation of SBC's hot cut processes, however, was in conjunction with SBC's Section 271 application to provide interLATA service within its service territory.19 Yet, the TRO states that any examination of the ILEC's existing hot cut processes in the various Section 271 proceedings are unreliable as an indicator of performance at the very high volumes at issue in this proceeding.
SBC's proposed processes anticipate no significant operational changes in underlying work functions, either in the central office, or in the LOC, or in the Local Service Center (LSC), where orders are received.20 SBC witness Heki, testified that SBC's central office workgroup has 2,372 central office technicians (i.e., the employees who do the inside work at the central office), a number sufficient to "keep up with increased demand."21 While each central office in isolation wouldn't have sufficient personnel to perform all hot cuts arising there, witness Heki testified that SBC could meet demand "by reallocating technicians or using overtime to address short term spikes in demand." 22 Thus, SBC argues, no new central office workers are required.
SBC claims that performing multiple hot cuts within a single batch will yield greater economies of scale, thereby facilitating large volumes of individual CLEC hot cut orders while lowering per line cost to convert end users to their circuit switched network.23 SBC points to new OSS offerings that will enhance the existing pre-ordering, ordering and provisioning OSS interfaces to provide CLECs more real-time information during the pre-ordering and provisioning phase, resulting in more automated processing for these requests.
SBC claims its proposed processes, designed to accommodate up to 200 cutovers per day per central office could transition the embedded base of UNE-P within the 27-month timeline laid out in the TRO even if every UNE-P in California migrated to UNE-L.24 SBC claims its processes accommodate more than sufficient volumes (1) for CLECs to meet the FCC timelines for transitioning the embedded base of UNE-P customers to stand-alone loops and (2) for SBC to meet the ongoing demand for hot cuts associated with CLEC acquisitions of new customers even as those volumes increase.
As of September 2003, SBC had approximately 1.26 million UNE-P lines in service spread across more than 600 central offices. Based on the 27-month timeline established by the FCC, and assuming the unbundling obligation for local switching was eliminated for all mass market customers in California, SBC computed that approximately 420,000 lines (an average of about 700 per wire center) would need to be cut by the 13th month, 840,000 lines by the 20th month, and the remaining portion of the 1.26 million by the 27th month pursuant to the FCC-prescribed timetable. To assess its ability to migrate the embedded base using its Defined Batch process, SBC focused first on its ability to migrate the embedded base in the one central office with the highest embedded base of UNE-P lines. At this central office, SBC estimates it would need to cut approximately 6,140 lines (at most) by the 13th month, 12,280 lines (at most) by the 20th month, and 18,421 lines (at most) by the 27th month in order to meet the FCC's transition timetable. Spread evenly over the 27-month transition period, this translates to about 32 lines per day (with 21 business days per month). If CLECs requested Saturday cuts, the daily number would be even lower.
SBC thus claims that its proposed volumes of up to 200 cutovers per day per central office more than accommodates the 32 lines per day that it would have to cut over in order to transition the embedded base within 27 months. Since the 32 lines per day is in the central office with the largest embedded base, SBC argues, its batch sizes can support migration of the embedded base consistent with the FCC's transition timetables. If 200 cutovers per day per central office were performed, the entire conversion for the two highest volume wire centers would be complete in approximately 90 business days - about one-ninth of the time the FCC allotted - and for SBC California's 307 lowest volume wire centers, the work would be completed in two business days.25
SBC further claims its processes will meet the anticipated demand associated with new CLEC acquisitions. To project that demand, SBC assumed that all current CLEC orders for UNE-P become orders for loops requiring hot cuts. As shown in Attachment 1.3 to Ex. 12 (Chapman), SBC estimates that the most activity of this sort that any SBC central office can expect is 88 additional cuts per day. Since no daily limit applies to orders for cutovers under the Enhanced Daily process, which is intended for new acquisitions, SBC claims that the process can comfortably meet the 88 cuts per day maximum projected for SBC's busiest central office. (See Ex. 12, at 23.)
B. Position of Other Parties
MCI and AT&T argue that SBC has underestimated the resources it will require to meet the increased hot cut demand assuming the elimination of
UNE-P. MCI argues that SBC should be required to modify and enhance its hot cut processes to provide the same timely, reliable, scalable and economically viable process as is currently available for UNE-P and the ILECs' retail services.26
MCI witnesses Starkey and Lichtenberg disagree with Chapman's claims concerning SBC's ability to accommodate increases in hot cut volumes. They argue that her analysis fails to recognize several factors that will impact the increased volumes in hot cut demand that SBC is likely to encounter. MCI identifies three factors in this regard: (1) transition volumes from existing UNE-P to UNE-L, (2) continued growth of CLEC lines as they achieve greater market share, and (3) churn related to existing customers moving between carriers. Starkey provided a model estimating the total number of hot cuts that SBC would likely be required to accommodate.
SBC projects, at a minimum, 197,000 hot cuts per month, over the
27-month implementation period prescribed in the TRO, assuming CLECs lost
access to UNE switching in all of the central offices being challenged by SBC.27 However, MCI witness Starkey projects that SBC would have to handle increasing monthly volumes of hot cuts over the entire 27-month period, and that SBC's maximum estimated volumes are at the low end of the level that SBC would have to handle in the first few months after UNE switching were removed.28 By the end of the 27-month period, MCI projects that SBC would have to handle 399,284 hot cuts per month, more than double SBC's monthly estimate of 197,000 hot cuts).29 MCI's projection of monthly hot cut demand is based upon a model that takes into account existing UNE-P lines, CLEC customer growth, and existing CLEC customer churn. MCI's model uses this information, in conjunction with the FCC's transition timeline for UNE-P to UNE-L conversion to estimate (1) the volume of hot cuts in any given month throughout the transition period and (2) the volume of hot cuts in the "steady state" once the transition is complete. MCI estimates that the "steady state" demand at the end of the transition period will actually be higher than demand experienced during any point during the transition period.30
C. Discussion
1. Batch Limits (47 C.F.R. § 51.319(d)(2)(ii)(A)(1))
The question of whether SBC's proposed volume limits are sufficient to meet CLEC demand is a function of what the demand for UNE-L turns out to be,
and the transition period over which UNE-P is transitioned to UNE-L. Because of the current uncertainty with respect to ultimate disposition of legal appeals of the DC Circuit Opinion, successor rules to the TRO, and potential renegotiation of interconnection agreements, we cannot determine precisely the nature, extent, pace, or duration of the transition from UNE-P to UNE-L in California. Thus, we can only render an assessment concerning the ability of SBC to meet hot cut demand under the assumptions made by parties in their testimony, recognizing that subsequent developments may result in different requirements as to the extent, timing, and schedule for UNE-P to UNE-L migrations.
SBC has offered a range of volume limits to be included in each hot cut batch per central office depending on the service option involved, as summarized above. SBC would permit every CLEC to obtain up to 50 Daily Batch cuts, plus up to 100 Defined Batch cuts, plus a negotiated number of Bulk Project cuts in every SBC central office every day (except Sunday and, for Daily Batch cuts, Saturday). SBC's hot cut processes provide for up to 100 loops per CLEC per day per wire center.31 SBC claims the only limitation on the number of cuts it can perform is the number of technicians that can work safely and efficiently on the same main distribution frame (MDF) at the same time. SBC's assumptions are based on a volume of about 200,000 hot cuts per month over the 27-month transition period outlined in the TRO.
AT&T concludes that in theory, SBC's proposed batch sizes could be adequate to meet demand, but questions whether SBC can meet actual hot cut demand without adequate testing and validation of the proposed processes. MCI, however, claims that SBC's proposed 100-loops-per-day volume limit is
definitely inadequate because there is no guarantee that the CLECs will be able to find a slot during which they can cut 100 loops, nor that a quick date will necessarily be available. MCI also claims that SBC's existing processes can handle only about one-half of SBC's own projected volume of hot cuts.
As noted above, the question of what constitutes an acceptable batch size depends on expected CLEC demand for hot cuts and the period over which the UNE-L transition is to be accomplished. CLEC demand for hot cuts is, in turn, a function of how extensively UNE-P is to be eliminated and replaced with UNE-L. The ability of SBC to handle hot cut volumes, both on batched basis and on an ongoing individual basis for churn and new customer growth is a function of a number of variables that currently remain in flux. One key variable is the number of UNE-P lines (both embedded and from ongoing churn) that would actually need to be cut over to UNE-L. Another variable is the actual timeline for beginning and completing the conversion. The TRO adopted a batch cut conversion schedule to be set in motion by state-mandated findings concerning mass market triggers. But because the TRO rules concerning mass market triggers have now been vacated and the FCC is in the process of preparing revised rules, the timing and extent of UNE-P conversions to UNE-L remain uncertain.
SBC's proposed batch size of 100 loops per day per CLEC per central office might be adequate depending on how the unknown variables are ultimately decided by subsequent FCC rules and/or by contractual agreements entered into by carriers. In the absence of this information, we cannot make definitive findings concerning a specific minimum batch size that must be accommodated. Accordingly, SBC's proposed batch size of 100 loops per day per CLEC per central office is hereby adopted only on an interim basis subject to further evaluation of performance metrics and testing to ascertain that this minimum batch size will be sufficient on an ongoing basis to meet CLEC hot cut demand on a seamless basis with the elimination of UNE-P.
As a cautionary matter, however, we believe that AT&T and MCI raise valid concerns about the potential for SBC's batch size to accommodate both the embedded base of UNE-P lines, as well as the ongoing churn from new lines, depending upon what assumptions are made about CLEC customer growth over the transition period. MCI assumes complete elimination of UNE-P statewide with transition required within the 27-month period prescribed by the TRO. To the extent that the elimination is more limited, or transitions over a longer period, MCI's assumptions would be overly pessimistic.
The essential concern with approval of a minimum batch cut volume would be to assure that no retail customer experienced any significant interruption in service as a result of the hot cut process, and that a "seamless" cut over could be achieved. We believe that appropriate performance metrics and incentives could be used to identify any potential issues with respect to hot cut performance so that remedial measures could be taken. These remedial measures might include arranging for additional workforce to handle higher hot cut demand and/or extending the schedule for the cut over of the embedded base to be completed.
2. Capability to Accommodate Hot Cut Demand and Ability to Meet Additional Demand (47 C.F.R. § 51.319(d)(2)(ii)(A)(3))
As a framework for analyzing SBC's capability to meet additional hot cut demand, we shall reference the assumptions that underlie the TRO, since that is what parties assumed in their analysis. To the extent that the actual UNE-P conversion scope is more limited, or the time for its completion is longer than
parties have assumed, SBC may be better able to meet the demand for hot cut volumes without service disruptions to affected CLEC customers.
We conclude that SBC's estimate of its ability to meet hot cut demand that would result from full elimination of UNE-P has not taken into account all of the relevant variables involved. Accordingly, before a final determination can be made concerning scalability to meet such demand, SBC needs to revise its forecasts of expected hot cut volumes to take into account the factors identified below. SBC also needs to maintain the capability and willingness to augment its workforce to meet the increased hot cut demand in a sufficiently timely and efficient manner to avoid service interruptions.
We are not persuaded that SBC's current workforce can meet its estimated volume of 197,000 hot cuts per month. SBC Witness Heki estimated that with 100% of SBC's "LFO-IN" workgroup dedicated to performing hot cuts, SBC could handle only an average of 100,000 hot cuts per month.32 Thus, SBC's own estimate indicates that the current LFO-IN workforce could handle only about half of the hot cuts required if CLECs lost access to UNE switching in all of the central offices that SBC is challenging. SBC would thus be unable to move the remaining half of CLEC customers' loops from SBC switches to CLEC switches
on a timely basis, leaving such customers without service, or forcing them to switch to SBC.
Heki assumed that the LFO-IN workforce could handle the increased demand based on the increase projected by witness Chapman for only a single day, but could not confirm whether such workforce reallocation and/or overtime could be sustained over the entire transition period assumed to be 27 months. SBC's prior experience with minor, short-term spikes in demand is not necessarily indicative of what would be expected with an ongoing required hot cut workload.33 The potential risk of failure to meet the increased demand is further illustrated by the fact that hot cuts currently constitute only about 5% of the work of a frame technician.34
Witness Heki did not know how many SBC central offices were manned and unmanned, nor the LFO-IN workforce percentages assigned to a specific central office versus how many are "roving." She did not know the maximum distance that an LFO-IN technician could be dispatched to assist with increased demand. It is also unclear how routine it is for SBC to reallocate LFO-IN technicians to meet demand spikes. Heki had not personally dispatched, nor was she aware of anyone who had dispatched an LFO-IN technician to cover an increase in workload at another central office in the two months preceding the date of her testimony.35 Heki failed to incorporate estimates for the work time
required for LFO-IN technicians to perform "throwbacks,"36 troubleshooting and repair, and loop "pass-overs"37 when conversation is on the line at the time appointed for a hot cut.
SBC's volume estimates did not take into account the number of hot cuts that will arise due to churn (customers moving between carrier), nor did they account for different categories of migrations that require hot cuts and/or wiring work on the frame. SBC did not take into account the tasks associated with hot cuts that the LFO-IN workgroup performs.38
Even if SBC's workforce grew to handle the increased volume of hot cuts, physical space constraints limit the number of technicians that can simultaneously perform wiring work efficiently and safely on the distribution frame,39 particularly since hot cuts for a large group of residential customers will generally appear at random frame locations.40 Technicians' ability to move around the distribution frame to make hot cuts is limited by: (1) the distribution frame size; 41 (2) the narrow crowded aisles between frames;42 (3) need for a limited number of sliding ladders.43 Thus, practical limits will remain on the number of technicians who can do simultaneous wiring work on the frame, regardless of the number of technicians that could be dispatched to a central office with high hot cut demand, without disrupting one another's wiring work, reducing efficiency and possibly creating safety hazards.44 SBC witness Mitchell admitted that current floor space plans do not anticipate TRO requirements. Mitchell foresees that SBC will evaluate growth needs after the Commission issues an order in this proceeding. Construction of new or expanded frames, however, will take three-to-six months.45 Thus, we question to what extent SBC could enhance its manual wiring process sufficiently to handle the increased volume of hot cuts.46
SBC witness Cusolito calculated the number of labor hours available to handle hot cuts in the existing workforce in the LOC, assuming that each worker would have 7-½ productive hours per day. That assumption accounted only for two 15-minute breaks for each worker per day, and did not deduct non-productive time such as vacations, sick days, holidays, personal days, training, company meetings or administrative tasks. During cross-examination, Cusolito acknowledged that his productive hour estimate per employee is too high.47
MCI argues that reasonable estimates for non-productive time, similar to the
estimate of Ms. Hernandez, the SBC expert for the LSC, would have yielded a much lower daily productive hour estimate of 6 hours per day rather than the 7 ½ hours Mr. Cusolito used in his calculations.48
AT&T witness Van De Water calculated that SBC would have to provision UNE-L at more than 45 times its current rate if UNE-P switching is no longer available. SBC has failed to evaluate whether main distribution frame (MDF) and independent distribution frame (IDF) capacity would need to increase to accommodate BHC migrations. Van De Water argues that added capacity is required due to added CLEC cabling that will be needed between the CLEC's collocation cages and the appropriate IDF or MDF.49 In view of these various uncertainties, as outlined above, we remain unconvinced that SBC could scale its hot cut processes to meet the increased the hot cut demand assuming full elimination of UNE-P on the assumed TRO timeline without expanding its workforce. In view of the uncertainties over SBC's capabilities to meet hot cut demand, we conclude that provision should be made for testing and validation to provide additional assurances that there would be no delays or other disruptions as a result of the transition from UNE-P to UNE-L. We address testing and validation requirements in Section XV below.
3. The BHC Process: Provisioning Interval Notice (47 C.F.R. § 51.319(d)(2)(ii)(A)(2))
Parties were in dispute concerning the appropriate provisioning interval for hot cuts. SBC incorporates a lag time of 13 business days to process CLEC orders in its "defined" batch process. SBC gives two reasons for the lag time:
(1) to enable SBC to reallocate personnel to handle the increased volume of hot cuts and (2) to schedule posting requirements for LFO-IN workers, who are unionized.
MCI claims that SBC's proposed interval is too long, and will create inconveniences for customers and a lack of transparency in the hot cut process.50 SBC claimed that it must have 13 business days' notice due to a provision in its labor contracts requiring advanced posting of employee work schedules.51 MCI claims that SBC does not need 13 business days to meet the schedule posting requirements of its union workers because the contract cited by SBC contains an exception to the schedule-posting requirement. For employees working the same hours between 9:00 a.m. and 5:00 p.m. Monday through Friday, no such posting is required.52 Thus, we agree that SBC should be able to offer CLECs an option to have their orders processed with only seven calendar days notice so long as the order specifies a hot cut between 9:00 a.m. and 5:00 p.m. week days.
This type of hot cut could be handled by LFO-IN personnel who would not be subject to the schedule posting requirement. Ms. Heki, the LFO-IN workgroup expert testified that if SBC did not have to comply with the schedule posting requirement, it likely could shorten the lag time for processing hot cut orders.53
In comments on the PD, SBC argues that it will be forced to incur extra overtime costs if it is forced to reduce its provisioning interval from 13 business days to a seven-day interval. We are not persuaded that reducing the provisioning interval necessarily results in any increased overtime costs. Since we are permitting SBC to limit the order processing duration to only regular business hours (i.e., 9 am to 5 pm week days), there would be no overtime pay periods involved in the processing of orders with a seven-day provisioning interval. Thus, we find no reason that SBC would necessarily have to incur overtime pay under a shorter provisioning interval given the restriction that processing of CLEC orders would be limited only to regular working hours.
We conclude that the time lag has a bearing on the cut over of embedded UNE-P customer migration because of the various dialing features such as call forwarding, call waiting and speed dial provided by the carrier's switch, as noted by AT&T.54 Thus, when customers are to be migrated from the SBC switch to a CLEC switch, customers must be notified, and must reprogram their phone sets. With a long delay, it would be more difficult for customers to remember when to reprogram their phone sets and/or CLECs will have to remind customers immediately before the hot cut date. Also, once an embedded UNE-P customer is told that a service-affecting event will take place on a particular day, they will expect the event to take place as scheduled. We thus find the proposal for the seven-day notice option to be reasonable and workable, and conclude that SBC's hot cut process should be modified accordingly.
15 FCC Rule 319(d)(2)(iv)(A), entitled "Transition timeline," required each CLEC to submit orders to transition its embedded base within a specified period after the ILEC was no longer required to provide unbundled local switching to serve mass market customers in a market. Specifically, each CLEC was to submit orders to transition one third of its embedded base of customers by the thirteenth month after such a Commission determination, the second third by the twentieth month, and the last third by the twenty-seventh month.
16 Tr. Ex. 31, Joint Testimony of Cusolito, et al., 12/15/03, 3.
17 Ex. 12, Chapman, pp. 18-25.
18 Ex. 50 (Hopfinger 12/12 Direct), at 19-20.
19 Id.
20 Ex. 31 (SBC Panel), at 13.
21 Ex. 31 (SBC Panel 12/15 Direct), at 42-43.
22 Id.
23 Ex. 31 (SBC Panel), at 13.
24 SBC's analysis assumes that existing UNE-L order volumes will continue, and focuses on the additional demand for hot cuts result due to migrating the embedded base and processing incremental UNE-L orders that would currently be UNE-P orders. (See Ex. 12 (Chapman), at 19.)
25 Because after September 2003, the embedded base grew at about 1% per month., (Jan. 28, 2004, Tr. at 8137-38; Jan. 30, 2004, Tr. at 8405) SBC acknowledges that these numbers that are a function of the total embedded base are slightly understated, but not enough to affect its conclusion. SBC argues that (1) the understatement is small; (2) SBC has built margins into its proposal; and (3) the understatement is more than outweighed by the fact that SBC will not be transitioning its entire embedded base. (See Id. at 8406-07.)
26 Ex. 143 (Lichtenberg/Starkey 1/15 Reply), at 6.
27 Tr. 2/3/04, (Cusolito), at 8859.
28 Ex. 143 (Lichtenberg/Starkey 1/15 Reply), at 30.
29 Id.
30 Ex. 143C (Lichtenberg/Starkey) 1/15 Reply) at 31-32.
31 SBC Brief at 86.
32 The 100,000 volume figure is determined in the following manner: The average volume of hot cuts that the LFO-IN workgroup can handle is approximately 5,000 per month, utilizing approximately 5% of its workforce. Thus, utilization of 100% of SBC's LFO-IN workforce dedicated to performing hot cuts represents a 20-times increase in the 5% workforce utilization. Assuming no other changes, this increase in workforce utilization would also equate to a 20-times increase in hot cut volume. Thus, the 5,000 per month volume multiplied by a factor of 20 would translate into an average volume of 100,000 hot cuts per month (i.e., 5,000 hot cuts * 20). (See Tr. 2/3/04 (Heki) at 8862).
33 Ex. 143 (Lichtenberg/Starkey 1/15 Reply), at 34.
34 Ex. 31, Cusolito et al. Joint Testimony, p. 42.
35 Tr. 2/2/04 (Heki) at 8715-8716.
36 If a problem develops during the hot cut process, SBC employs a "throwback" process in which the customer's loop will be reconnected to SBC's switch.
37 The technician will expend additional time beyond a normal hot cut when there is conversation on the line because he or she must "pass over" that loop to the next cut, then return and retest the loop to determine if there is conversation on it.
38 Ex. 143 (Lichtenberg/Starkey 1/15 Reply), at 27-30.
39 Tr. 2/3/04, (Heki), at 8828; Ex. 16 (SBC panel 11/7 direct), at 22.
40 Ex. 142 (Starkey 12/15 Direct), at 26.
41 Tr. 2/3/04, (Heki), at 8824-8825.
42 Ex. 33C, (see section labeled "environment").
43 Tr. 2/3/04, (Heki), at 8825-8826.
44 Tr. 2/3/04, (Heki), at 8828.
45 Tr. Vol. 58, 2/4/04; 8959
46 Ex. 143 (Lichtenberg/Starkey 1/15 Reply), at 4, 27.
47 Tr. 2/2/04 (Cusolito), at 8737-8738.
48 Tr. 2/2/04 (Hernandez), at 8731.
49 Ex. 154C, Van de Water Testimony, p. 45
50 Ex. 143 (Lichtenberg/Starkey 1/15 Reply), at 60-61.
51 Ex. 31 (SBC panel 12/15 Direct), at 16.
52 Ex. 35, at 98 (Section A1.03 B.2.).
53 Tr. 2/3/04, (Heki), at 8834:13-16.
54 AT&T Opening Brief, at 188 n.598.
